Social Protection, Growth and Employment by United Nations Development Programme

Empowered lives. Resilient nations.

Social Protection, Growth and Employment Evidence from India, Kenya, Malawi, Mexico and Tajikistan

May 2013

First published by the United Nations Development Programme in the USA in 2013.

9-789211-263633

Cover design by Saswata Alexander Majumder Cover Image by Bengy/Shutterstock For a list of publications please contact: United Nations Development Programme Bureau for Development Policy One United Nations Plaza New York, NY 10017, USA Tel: +1 212 906 5081 Email: poverty.reduction@undp.org website: www.undp.org/poverty

Disclaimer The views expressed in this publication are those of the authors and do not necessarily represent those of the institutions to which they are affiliated or the United Nations or their Member States, including UNDP.

Foreword We live in times that seem to be defined by shocks and vulnerabilities, and these shocks and vulnerabilities have the potential to slow down, or even reverse, achievements in human development. The recent global economic downturn has already had some negative effects on human development indicators, for example. In recent decades, there has been a renewed interest in social protection policies as a way to help households cope with the negative effects of economic policies. Many countries have adopted cash transfer programmes, for example, as a way to support the income of poor households and encourage them to invest more in developing human capital through education and health. While evaluations of these programmes show measurable improvements in some human development indicators such as school attendance and health status, less is known about their sustained effects on poverty reduction and inclusive growth. This sustained effect can be potentiated if the positive effects on education and health, for instance, are accompanied by a better possibility of beneficiaries to insert themselves productively in the labour market and the local economy. Employment is a key strategy for sustainable poverty reduction and promotion of inclusive growth. However, economic growth, even when rapid, has not always created jobs, particularly jobs for the poor. Furthermore, in the present global context, growth has slowed down and employment has suffered as well. Even when growth recovers, the need for innovation and re-reorganization of production and service delivery in order to reduce costs might well mean a weaker demand for labour. Less rapid growth and lower employment intensity of growth will make inclusive growth and progress in human development more difficult. In addition to this, the quality of the jobs created also matters in terms of human development outcomes. Insecure, unstable and unprotected employment has limited potential to have a lasting effect on poverty reduction. In this sense, a closer linkage of employment policies with social protection policies and programmes can also promote greater resilience both at the level of the household, but also the economy as a whole. Both employment and social protection thus are keys to inclusive growth, sustained poverty reduction and improvement in human development. However, establishing the linkages between them so as to potentiate the positive effects of employment and social protection remains an area where much needs to be learned. The present volume addresses this question by looking at evidence from six country case studies to assess how to enhance the employment impact of social protection programmes by improving the consistency of objectives between them and the macro, sectoral and territorial policies adopted by countries. These studies look at a range of social programmes to assess their potential and limitations for supporting more inclusive growth and for sustained poverty reduction. They also focus on design and implementation questions as well as their links with other social and macroeconomic policies. This publication offers valuable insights into these issues and highlights key policy considerations for policy makers. It also offers an important

contribution to the discussion on how to think about the linkages, which requires further research and work. I hope that the present volume would be useful to academics, researches, policy-makers and broader development practitioners.

Selim Jahan Poverty Practice Director Bureau for Development Policy United Nations Development Programme

Table of Contents Foreword

iii

Boxes, Figures and Tables

List of Acronyms and abbreviations

List of Contributors

xii

Acknowledgements

xiii

1. Growth, Employment, Poverty and Social Protection: A conceptual framework Albert Berry

2. Social Protection and Labour Markets in Malawi: the Centrality of Agriculture Andrew Charman

3. Addressing Transitory and Chronic Poverty: Social Policies in Peru 1990–2010 Janina Leon

4. Challenges of transition economies: Economic reforms, emigration and employment in Tajikistan Omair Amir and Albert Berry

157

5. Labour, Income and Social Programmes in Contemporary Mexico Carlos Salas

201

6. Guaranteeing Jobs for the Rural Poor: An Assessment of India’s MGNREGA public works programme Eduardo Zepeda, Diana Alarcon, Scott McDonald, Chandan Sapkota, Manoj Panda, and Ganesh Kumar

235

7. The Youth Employment Challenge: The insertion of Kenya’s youth into formal and informal jobs Eduardo Zepeda, Fatou Leigh, Lydia Ndirangu and Jacob Omolo

257

8. Policy, Institutional and Methodological lessons from six country studies Albert Berry

287

Boxes, Figures and Tables Boxes Box 3.1: The new Humala government and its social policy agenda—a quick glace

130

Box 3.2: The new Humala government and proposals for microenterprises and MFIs— 135 a quick glance Figures Figure 2.1: Population Age Structure, 2008

Figure 2.2: Maize Production, Disaggregated by Variety, 1987–2010

Figure 2.3: Burley Production and Yield, 1987–2010

Figure 2.4: Cassava and Sweet Potato Production, 1987–2010

Figure 4.1: FAO Food Price Index in Real Terms, 1961–2010

172

Figure 4.2: Food Price Trends in Tajikistan and Globally, 2007–2010

173

Figure 4.3: Trends in Remittances, 2008–2011

173

Figure 4.4: Productivity of Land by Farm Type, 1991–2006

179

Figure 4.5: Cropping Pattern in Household Plots and Dehkan Farms (DF), 2007–2008

179

Figure 5.1: Female Labour Force Participation Rates, by Age, 1995 and 2010

207

Figure 5.2: Open Unemployment Rate, 2000–2011

208

Figure 7.1: Population Trends by Age Group and Area of Residence, 1955–2100

259

Figure 7.2: Youth and Adult Unemployment Rates, 2005/06

260

Figure 7.3: Activity Choices of the Working-age Population 2005/06

260

Figure 7.4: Youth Unemployment Rates and Activity Choice by Area, 2005/06

262

Figure 7.5: Youth Unemployment Rates and Activity Choice by Gender, 2005/06

263

Figure 7.6:Youth Unemployment Rates and Activity Choice by Educational Attainment, 2005/06

264

Figure 7.7: Youth Unemployment Rates and Activity Choice by Gender, Area of Residence and Age, 2005/06

265

Figure 7.8: Unemployment Rate by Gender, Type of Area of Residence and Education, 2005/06

266

Figure 7.9: Employment and Unemployment by Age, Type of Area of Residence, Gender and Education, 2005/06

267

Figure 7.10: Wage Variation by Age in Rural and Urban Areas, 2005/06

270

Figure 7.11:Wages by Type of Employer and Location, 2005 /06

270

Figure 7.12: Wages by Gender, 2005/06

271

Figure 7.13: Wages by Type of Employment for Young People and Adults by Gender, 2005/06

272

Figure 7.14: Wages by Education, Age and Type of Work, 2005/06

272

Figure 7.15: Number of Young People Employed in KKV Public Works by Ministry

277

Tables Table 1.1: Different Types of Provision of Direct Social Protection against Emergencies

Table 1.2: Links between Social Protection Components and Employment, and Relations to Motors of Pro-poor Growth

Table 1.3: Key Considerations around Selected Social Protection Policies

Table 1.A1: Asian Development Bank’s Components and Subcomponents of Social Protection

Table 2.1: Indicative Features of Poverty

Table 2.2: Proportion of Households That Own Selected Assets, 2004 vs. 2008

Table 2.3: Proportion of Households Affected by Severe Shocks, c. 2004

Table 2.4: Timeline of Key Institutional Actions for Social Protection

Table 2.5: Social Protection Interventions, Estimated Beneficiaries and Accumulated Costs, c. 1994–2010

Table 2.7: Cost–Benefit Comparative Analysis: Starter Pack, AISP and Food Aid

Table 2.8: Maize and Tobacco Labour Input

Table 2.9: Distribution of Benefit from Productivity Enhancement Measures

Table 3.1: Government, Social and Basic Social Expenditure, 1999–2010

109

Table 3.2: Total Government Expenditure and Social Expenditure, 1999–2010

110

Table 3.3: Basic Social Expenditures by Type of Expenditure, 1999–2010

110

Table 3.4: Head of Household by Social Programme and Characteristics, 2000, 2005 and 2009

112

Table 4.1: Sources of Cash Income, Tajikistani Households, 2006

169

Table 2.6: Maize Fertilizer Consumption, 1996–2007

Table 4.2: Estimates of the Share of Individual and Corporate Sector in Land, Based on Data from Goskomstat and the State Agency for Surveying, Cartography and Land Use, 2006

178

Table 5.1: Macroeconomic Data on Mexico, 1980–2011

202

Table 5.2: Poverty Data for Mexico, 1984–2010

203

Table 5.3: Latin America’s Loss of Market Share in the USA to China, 1996–2009

206

Table 5.4: Labour Force Participation Rates, by Gender, 1995–2010, 2nd quarter

207

Table 5.5: Increases in Employment, by Size of Establishment, 2006–2010, 2nd quarter

209

Table 5.6: Percentage of Employment by Job Position, by Economic Sector, 1995 and 2010, 2nd quarter

211

Table 5.7: Employment by Economic Sector, by Size of Establishment and by Whether Paid or Not, 2010, 2nd quarter

212

Table 5.8: Percentage of Paid Workers with Social Security Benefits, by Economic Sector, 2010, 2nd quarter

213

Table 5.9: Percentage of Paid Workers with Labour Benefits, by Economic Sector, 2010, 2nd quarter

213

Table 5.10: Percentage of Paid Workers with Labour Benefits, by Type of Labour Contract, 2010, 2nd quarter Table 5.11: Contractual Arrangement of Paid Workers, by Economic Sector and Size of Establishment, 2010, 2nd quarter

214

215

Table 5.12: Average Monthly Labour Income by Economic Sector, Size of Establishment and Job Position, 2010, 2nd quarter

217

Table 5.13: Number of Microbusinesses by Whether the Owner Hires Paid Workers and by Number of Workers, 2008

219

Table 5.14: Main Reasons for Beginning a Small Business, 2008

220

Table 5.15: Microbusinesses, by Economic Activity, by Presence of Paid Workers or not, and by Age, 2008

220

Table 5.16: Main Problems Reported by Microenterprises, by Availability of Fixed Location for Business, 2008

221

Table 5.17: Type of Registration with an Institution by Microenterprises, 2008

222

Table 5.18: Trade Union Membership by Type and by Type of Microenterprise, 2008

222

Table 5.19: Access to Health Services, by Type of Microenterprise and Type of Health Service, 2008

223

Table 5.20: Flows between Employment, Unemployment and Economic Activity, 2010, 2nd and 3rd quarters

224

Table 5.21: Distribution of Employed Persons by Size of Establishment, 2010, 2nd and 3rd quarter

224

Table 5.22: Support Programmes for Micro-units, Known of and Used, 2008

225

Table 5.23: Access to Community Health Insurance, by Whether Employed or Not and, for Those Employed, by Decile of Per Capita Family Income, 2010, 2nd quarter

226

Table 5.24: Coverage of the Employment Support Programme (PAE), 2003–2010

227

Table 6.1: NREGA’s Job Creation and Household Coverage, 2006–2011

242

Table 6.2: Share in the Total Number of Person-days Worked on NREGA Projects by Women, SCs and STs, 2006/07–2011/12

242

Table 6.3: Distribution of NREGA Works by Project, 2006/07 and 2011/12

243

Table 6.4: Central Government Expenditures on NREGA Programmes, 2006/7–2011/12

244

Table 7.1: Typology of Kenya’s Employment Creation Interventions, 1963–2011

275

List of Acronyms and abbreviations AAH

Action Against Hunger

ADB

Asian Development Bank

ADMARC

Agricultural Development and Marketing Corporation

AISP

Agricultural Inputs Subsidy Programme

ALDSAP

Agricultural and Livestock Development Strategy and Action Plan

ATU

A trabajar urbano

CCT

Conditional cash transfer

CIS

Commonwealth of Independent States

DEVPOL

Development Policies

DFID

Department For International Development (UK)

ENAHO

National Survey of Households

ENAMIN

National Survey of Microenterprises

ENE

National Employment Survey

ENOE

National Survey of Occupation and Employment

European Union

FDI

Foreign Direct Investment

FFW

Food for work

Fondo PyME Small and Medium Sized Enterprises Fund GDP

Gross domestic product

GNI

Gross national income

GoM

Government of Malawi

IDB

Inter-American Development Bank

IHS

Integrated Household Survey

ILO

International Labour Organization

IMF

International Monetary Fund

IMSS

Mexican Social Security Institute

INEGI

National Institute of Statistics and Geography

INGO

International Non-Governmental Organizations

ISSSTE

Institute for Social Security and Services for State Workers

Phosphorous

KIHBS

Kenya Integrated Household Budget Survey

KKV

Kazi Kwa Vijana

KShs

Kenyan Shillings

KYEP

Kenya Youth Empowerment Project

LAC

Latin America and Caribbean

LDC

Least Developed Country

LDW

Local Development Window

Moving average

MASAF

Malawi Social Action Fund

MCP

Malawi Congress Party

MDG

Millennium Development Goal

MFI

Microfinance institution

MoAFS

Ministry of Agriculture and Food Security

MOYAS

Ministry of Youth Affairs and Sports

MPRS

Malawi Poverty Reduction Strategy

MPVA

Malawi Poverty and Vulnerability Assessment

MSME

Micro, small and medium-sized enterprise

MTC

Ministry of Transport and Communication

MVAC

Malawi Vulnerability Assessment Committee

MyPE

Micro y PequeĂąa Empresa (micro and small enterprise)

Nitrogen

NAFTA

North American Free Trade Agreement

NGO

Non-governmental organization

NREGA

National Rural Employment Guarantee Act

NSO

National Statistics Office

NYS

National Youth Service

Organizacion de base (civil society organization)

OBCs

Other backward classes

OECD

Organisation for Economic Co-operation and Development

OPM

Office of the Prime Minister

OPV

Open-pollinated varieties (maize)

Potassium

PAE

Employment Support Programme

PCM

Presidency of Ministersâ&#x20AC;&#x2122; Council

PEN

Peruvian Nuevos Soles (currency)

PET

Temporary Employment Programme

PRONAFIM

National Programme for Microenterprise Financing

PVD

Provias Descentralizado

PVN

Provias Nacional

PWP

Public works programme

Sulphur

SACA

Smallholder Agricultural Credit Administration

SAL

Structural Adjustment Loans

SAM

Social Accounting Matrix

Scheduled castes

SFFRFM

Smallholder Farmers Fertiliser Revolving Fund Mechanism

SIS

Sistema Integral de Salud (Integrated Health System)

SME

Small and medium-sized enterprise

SpyME

Sub-ministry for Small and Medium-sized Enterprise

SSR

Soviet Socialist Republic

STPS

Ministry of Labour and Social Provision

STs

Scheduled tribes

TIP

Targeted Inputs Programme

TJS

Tajikistani Somoni (currency)

TSC

Teachers Service Commission

UDF

United Democratic Front

UNDP

United Nations Development Programme

USAID

United States Agency for International Development

WFP

World Food Programme

WMS

Welfare Monitoring Survey

List of Contributors Diana Alarcón is Senior Economic Affairs Officer at the Development Policy and Analysis Division of the U.N. Dept. of Economic and Social Affairs. Formerly Leader of the Inclusive Development Cluster. Omair Amir is an Independent Researcher with a focus on Central and South Asia. Previously he was Research Assistant at the Department of Economics at the University of Toronto. Albert Berry is Professor in the Department of Economics at the University of Toronto. Andrew Charman is the Director of the Sustainable Livelihoods Foundation in South Africa. Ganesh Kumar is Professor at the Indira Gandhi Institute of Development Research, Mumbai, India. Fatou Leigh is Economic Advisor in UNDP Kenya. Janina V. Leon, Ph.D. is Professor in the Department of Economics at the Pontificia Universidad Catolica del Peru. Scott McDonald is Professor of Economics in the Department of Accounting, Finance and Economics at Oxford Brookes University, United Kingdom. Lydia Ndirangu, Ph.D. is Assistant Director of the Research Centre at the Kenya School of Monetary Studies, Kenya. Jacob Omolo, Ph.D. is Lecturer at the Kenyatta University, Kenya. Manoj Panda is director of the Institute of Economic Growth, New Delhi, India. He was previously director of the Center for Economic and Social Studies, Hyderabad and Professor at the Indiar Gandhi Institute of Development Research, Mumbai. Carlos Salas-Páez is Professor at CESIT-Instituto de Economía of UNICAMP, Brazil. Chandan Sapkota is Economics Officer at the Asian Development Bank, Nepal Resident Mission. He was a junior fellow at the Carnegie Endowment for International Peace from 2009 to 2010. Eduardo Zepeda is Inter-regional Policy Coordinator of the Development Policy and Analysis Division of the U.N. Dept. of Economic and Social Affairs. He was previously Senior Policy Advisor in the Poverty Practice of the Bureau for Development Policy of the United Nations Development Programme (UNDP) and Senior Associate in the Trade Equity and Development Program at the Carnegie Endowment for International Peace.

Acknowledgements This work would not have been possible without the leadership—both at the intellectual and management levels—of Selim Jahan, Director of the Poverty Practice of UNDP. His conceptual contributions and unfaltering support for this work led to the successful conclusion of a product based on sound research and analysis that can contribute to the policy debate about the linkages between social and economic policies. Professor Albert Berry provided the conceptual backbone of the project and coordinated the preparation of the case studies that are the heart of this volume. Dr. Berry was responsible for the development of the conceptual framework and the overview chapter. Beyond that, he acted as content editor of all country studies and in many cases was actively involved with authors in the preparation of chapters, making sure all parts effectively contributed to the whole. The project was launched by Diana Alarcón during her tenure as Leader of the Inclusive Development Cluster in the Poverty Practice. Her solid knowledge of the field of development economics and policy, effective management and indefatigable energy took this work from an idea to a viable research initiative. Eduardo Zepeda also had a key leadership role in the development of this project contributing substantively, not only to his own chapters, but to the overall arguments and evidence presented in the volume. In addition to managing the day-to-day activities of the project, Claudia Vinay also contributed substantively throughout the process. Her work could not have been possible, though, without the steadfast support of Mareike Eberz who became involved in the latter stages of the project. She provided able research support to finalize chapters and efficiently and cheerfully managed the complicated peer review process and a myriad other details. Almudena Fernández stepped in to help when the need arose to make sure that the process moved along — this on top of her other responsibilities. Eliane D’Pierre was the cornerstone on the operational side, efficiently managing contracts, payments, procurements, planning events, communications and travel needs. A note of thanks goes also to Anuradha Seth, Team Leader of the Inclusive Green Growth Team of the Poverty Practice, who patiently supported the team’s involvement in this work at the tail end when the small details seemed never-ending. Special thanks, of course, go to all authors of the background studies, which served as a basis for the preparation of the country chapters in this publication. They are: Omair Amir, Andrew Charman, Janina León and Carlos Salas. For the India chapter, the collaboration with the Carnegie Endowment for International Peace was instrumental and facilitated the participation of Ganesh Kumar, Scott McDonald, Manoj Panda, and Chandan Sapkota. For the Kenya chapter, special thanks go to the UNDP Kenya Country Office and Regional Bureau for Africa, in particular to Fatou Leigh for her contributions on both the substantive and coordination sides. Lydia Ndirangu and Jacob Omolo were also co-authors for the Kenya study and Moses Mathu provided research support.

Thanks also to all the peer reviewers both from within UNDP as well as from external institutions and to all the participants in the March 2011 technical meeting. They include: Tuya Altangerel (UNDP), Rosemary Atieno (University of Nairobi), Babken Babajanian (Overseas Development Institute), Cesar Patricio Bouillon (Inter-American Development Bank), Mariana Campos (Independent Consultant), Griet Cattaert (International Labour Organization), Mohamed Chemingui (Independent consultant), Ricardo Fuentes (OXFAM), Alan Gelb (Center for Global Development), Degol Hailu (UNDP), Radhika Lal (UNDP), the late Sebastian Levine (UNDP), Sheila Marnie (UNDP), Paola Pereznieto (Overseas Development Institute), George Perkovich (Carnegie Endowment for International Peace), Carola Pessino (Centre for Global Development), Robert Pollin (University of Massachusetts, Amherst) and, Fabio Veras Soares (UNDP). Lance Garmer copy-edited the Malawi case study and Yurik Sandino copy-edited a first version of the Kenya study. Jon Stacey of the Write Effect was in charge of copy-editing the rest of the chapters and of ensuring overall consistency of the text. Shams Banihani provided detailed quality assurance of the product and efficiently managed the final stages of the publication. Alex Majumder took care of the design and layout.

1 Growth, Employment, Poverty and Social Protection: A conceptual framework Albert Berry Introduction It can be argued that most reduction of income poverty in the developing world is due either to reasonably pro-poor economic growth or to social protection policies of one sort or another, with growth usually the main single factor at work, especially in lower-income countries. Thus the challenge associated with the design of good social protection policies begins with the nature of the growth process. A successful package of economic and social policies must produce (i) a good rate and pattern of economic growth to reduce poverty directly; (ii) a well-designed system of social protection to defend those still left in poverty despite the growth achieved; and (iii) internal consistency between the two broad categories of policy involved, such that neither cancels out the positive effects of the other.1 It is particularly crucial that social protection policy be as consistent as possible with the creation of good jobs, since it is mainly through job creation that growth contributes to poverty reduction. There are important differences in the design of pro-poor growth and effective social protection programmes across countries, some of them associated with a countryâ&#x20AC;&#x2122;s level of development. Although economic growth is almost always a key factor in poverty reduction, this is especially true for low-income, often mainly agricultural countries. Such countries are generally characterized by a low public spending capacity (due to low taxing capacity), a low implementation capacity and a higher presence of community mechanisms that can protect vulnerable people

Social protection, growth and employment

from certain types of economic crisis. Their public social protection policies must be consistent with the growth process, both because, ceteris paribus, the lower per capita income is, the more important growth is and because, given the necessarily limited role of public social protection, it is essential that it does not hinder but rather complement growth. As countries progress economically, this balance changes, to the point where in countries with high fiscal and implementation capacity, social protection can be a major tool against much poverty and vulnerability. Of particular interest are the social protection initiatives that overlap with labour market processes and hence affect labour outcomes. Such overlaps take different forms according to where a given policy is on a spectrum from pure protection (no impact on the person’s/family’s capacity to earn income) to pure income growth promotion such that the protection provided comes exclusively from the higher income. Some social protection interventions, though, may directly diminish income earned, such that the net benefit due to the programme is less than the transfer involved. A further relevant distinction is that between social protection whose income-creating effect occurs immediately—for example, employment-generating public works programmes—and those whose income-generating effect comes later, as with conditional cash transfers (CCTs) designed to keep the children of low-income families in school longer or school lunch programmes whose main impact may be a reduction of malnutrition and a healthier person over his/her lifetime. These distinctions by no means fully capture the causal interactions between social protection programmes and either personal/family income growth or the overall economic performance of an economy. Where poverty or extreme poverty is the result of a vicious circle and it is possible for an appropriate intervention to prevent that circle from occurring or to constitute the first step out of it, the social protection intervention involved can have a special payoff. In this context the way such family economic dynamics occur may be more important than the details of which interventions or elements of assistance have what impact (González de la Rocha et al., 2012). There is thus a need to integrate the analysis of individual social protection interventions with the dynamics of poverty as they play out in given settings, both for the family and for the country. Since employment provides the great bulk of income for lower-income families, the opportunities provided by the labour market and self-employment earnings are likely to be pivotal to the potential success of exit from poverty, sometimes set in motion or at least aided by one or more social protection programme. Taking a broad perspective on how family welfare may be improved leaves us with the possibility that some of the best policies to ‘protect’ people against bad economic outcomes are ones that promote pro-poor growth. Pro-poor growth is growth that raises poor people’s incomes by a greater percentage than those of the non-poor.2 It means that poor people’s incomes are rising both because of growth per se and because the pattern of that growth is pro-poor. In that case, growth objectives and social protection objectives may fully overlap in the sense that the best policies for the first objective are also the best ones for the second. At the other end of the spectrum, the policies that generate fast growth may be exclusive, such that many people’s needs are not satisfied by that growth process and may even be damaged by it. Sometimes even the most beneficial growth pattern that a country can manage excludes too many people from

A conceptual framework

the benefits and thus needs to be complemented by a strong social protection package designed to protect those not otherwise taken care of. A number of social protection programmes have been introduced or greatly expanded in developing countries over the last few decades, and there is now a large literature that assesses them (UNRISD, 2010; Hujo and McClanahan, 2009; Grosh et al, 2008) Mainly independently of the social protection literature there has been an ongoing debate on how satisfactorily labour markets and labour outcomes have been evolving and what policies and structural factors lie behind differences in that performance (Pages, 2004; Perry et al, 2007; World Bank, 2012). Finally, there is still a small literature that addresses aspects of the interface between social protection and labour outcomes and/or aggregate economic performance (Auerbach et al, 2007; Levy, 2008; Berry, 2013). This area, to which this volume attempts to contribute, is a very important one, partly because social protection policies do matter in an increasingly important way to many people in many countries and partly because there are a number of important possible causal links between labour market functioning/labour outcomes and those social protection policies. Any discussion of the benefits and costs of social protection must begin by recognizing the wide range of programmes, their differing objectives and the wide range of country settings. â&#x20AC;&#x2DC;Country settingâ&#x20AC;&#x2122; here includes economic structure, level of development and levels of poverty and related social ills, on the one hand, and administrative capacity of government or other relevant institutional actors, on the other. The heart of this volume consists of six case studies (Malawi, Tajikistan, India, Kenya, Peru and Mexico) involving a wide range of country settings for social protection and hence for the interfaces of those policies with economic growth and labour outcomes. They similarly feature a wide range of social protection policies, of total resources directed towards them and probably also of the overall impact of those policies on welfare, poverty, employment and other important variables. Malawi and Tajikistan have lower incomes, so growth remains central to long-term success, while Peru and Mexico have higher incomes and are using their greater social protection potential in a variety of ways. In Kenya a lack of jobs for young people is of special concern. India has recently launched the single biggest social protection programme ever, with great potential to reduce poverty and economic insecurity but also with enormous implementation challenges. The case study approach makes no attempt to reach wide-ranging generalizations about any specific types of social protection policy; rather, the aim is to throw light on a range of issues raised by the cases included. They are mostly direct policy issues, i.e. involving how well various types of policies have been or could be working. In some cases, especially the India study, there is heavy emphasis on methodology of analysis, in particular the importance and value of a general equilibrium approach to estimating programme benefits. This introductory chapter lays out a taxonomy of linkages between social protection and the economic variables towards which policy is directedâ&#x20AC;&#x201D;growth, employment, inequality, poverty

Social protection, growth and employment

and economic vulnerability. It aims to provide a framework for the sort of analysis carried out in the country case studies. Like those studies, it places special emphasis on the linkages between social protection, economic policy and growth, to highlight the importance to policy-makers of avoiding the too-common institutional separation between these policy areas. The final overview or comparative chapter summarizes the main results emerging from each of the case studies and draws some overarching conclusions.

Background concepts for thinking about growth, poverty reduction and social protection ‘Social protection’ is one of the ways to reduce and/or alleviate the impacts of poverty and people’s vulnerability to it. Before looking in detail at the key issues involved in the design of a set of policies that meet these criteria, it is helpful to review relevant background concepts. A country’s average income level and its growth broadly depend on: • resources available — natural resources such as land, physical capital, human capital (knowledge and skills); • technologies available and in use, which determine the potential productivity of the resources; • the degree of resource utilization (what share of available resources are actually being used); and • the efficiency of resource utilization, i.e. the degree to which the resources are being allocated to their most productive potential uses. The rate of growth of a country’s economy depends on changes over time in these four variables, i.e.: • the rate of investment in physical and human capital, and of discovery of new natural resources; • the rate of total factor productivity improvement (or technological change, in economic parlance);3 • changes in the degree to which available resources in the economy are utilized; and • changes in the effectiveness with which those resources are employed (allocated among alternative uses). The same four categories of factors are relevant to the growth of any economic unit, from the world economy to national, sub-national, local and on down to individual companies. All fast-growing economic units have high rates of investment; they typically also have high rates of productivity growth and usually improve the rate of resource utilization and sometimes the

A conceptual framework

effectiveness of resource allocation. Causal interactions among these four factors are also typically at play. Thus new technologies not only directly contribute to growth but also induce more investment than would otherwise have occurred. A high rate of investment tends to increase the degree of utilization of labour (i.e. it cuts unemployment or underemployment). The incidence of poverty (the share of people with income below a certain minimum absolute level, the ‘poverty line’) depends on the average income in the country and the level of inequality. For a given average income, the greater the level of inequality, the greater the share of people in poverty. Changes in the level of poverty reflect changes in average income and/or in its distribution. When growth occurs with no change in income distribution (i.e. all people or families experience the same percentage growth of income), a smaller share of the population remains in poverty. When inequality falls without any change in average income, more households are clustered in the middle ranges of income. Over recent decades, the great bulk of the poverty reduction that has occurred in the developing world has been due to growth, with few cases of inequality reduction (Amann et al, 2005; Berry and Serieux, 2007; Dollar and Kraay, 2002). But the obvious objective is pro-poor growth. Poverty and inequality have been defined in various ways, by the level and distribution of household income, with or without any transfers from the state; the level of private consumption; or the level of private plus public consumption (i.e. including publicly provided education and health). The most useful definition involves both private and publicly based incomes, i.e. transfers from the state or direct support (e.g. publicly provided health, education etc.).4 This perspective invites the distinction between ways to raise poorer people’s earned incomes through employment, and ways to improve their welfare through the tax and transfer system, which can provide education, health, welfare payments and other poverty-reducing forms of assistance, although these two categories frequently overlap. Practically speaking, employment policy is central to poverty alleviation, since the level and quality of employment is a basic determinant of national income and its distribution. The main source of income for 80–90 percent of families in most countries is their labour. Vulnerability, the possibility of being thrown into a state of emergency or distress by some uncontrollable event, is naturally related to poverty, since poor people are more vulnerable to almost all types of crises. Those who are better-off economically can insure themselves against certain types of risk—for example, by saving during their working years to prepare for retirement or by buying insurance against unlikely events. Poverty and vulnerability diminish or threaten people’s welfare, either on a continuing basis (poverty) or through occasional emergencies and the psychological toll extracted by the awareness of and worry about that vulnerability (Naudé et al., 2009; Naudé and Santos-Paulino, 2010). Who is poor (and hence more vulnerable) depends both on how the economy itself functions to generate the ‘primary distribution of income’ and how government or other redistributive policies or practices then affect the ‘secondary (or final) distribution of income’. The primary income distribution, which reflects both the economic behaviour of individuals and the

Social protection, growth and employment

economic context in which they live, determines who will suffer from poverty in the absence of any societal response to prevent them from doing so. As for inequality, the main and most obvious causal factor is the ownership pattern of the inputs to production, i.e. natural resources, physical capital and human capital or skills. Concentrated ownership of these inputs guarantees a concentrated distribution of primary income. Also important, however, are the technologies in use; thus a country with a large population but which employs a lot of modern, capital-intensive technology (e.g. large tractors on its farms) is more likely to have an unequal distribution of income, whereas one that uses technologies more in keeping with the available resources will suffer less inequality. Both growth and inequality are influenced by many aspects of the structure of the economy and the institutions of the society. Among the relevant structural features are the composition of the economy by sector (agriculture, manufacturing, transportation etc.), by size of firm and by technology used. A well-recognized source of inequality in many developing countries is the high level of dualism, whereby much of the economyâ&#x20AC;&#x2122;s capital is employed in capital-intensive (usually modern) technologies that employ few workers, thus leaving most of the labour force to operate with very labour-intensive technologies, usually in small farms or the informal nonagricultural sector.

Social protection The levels of per capita income and of inequality that result from the functioning of the market economy always leave some people in poverty and/or vulnerable to emergencies; social protection policies are designed to alleviate these problems (and, where possible, to contribute to pro-poor growth). When considering how the attempt to reduce poverty can be made most consistent with the attempt to generate fast pro-poor growth, it is helpful to highlight a number of distinctions within the broad area of social protection. Prevention of poverty vs. prevention of specific emergencies: Social protection can respond to two types of emergency. In one, the basic problem is low income, so the assistance is directed at raising income. This category includes welfare programmes (that provide income to the beneficiaries, either monetary or in kind) and employment guarantee systems (employment of last resort); they are anti-poverty programmes in the most direct sense. They may be permanent, when the problem is chronic poverty, or temporary to address temporary (e.g. seasonal) poverty. The other variant is directed at more narrowly defined emergencies, especially ones involving health, but also such problems as vulnerability to food shortage or to abusive treatment. Prevention vs. reaction: People can be protected against social and economic emergencies in a range of ways. Table 1.1 summarizes the forms that direct social protection can take and the ways it may be provided.

A conceptual framework

Table 1.1: Different Types of Provision of Direct Social Protection against Emergencies Form of protection

Context

Provider of resources

High-income protection

The individual or family has enough income to deal with a situation that would constitute an emergency for a lower-income family.

Individual or family

Preventive protection (protection through pre-planning)

Because of the high costs involved, advance savings are needed to provide a fund to deal with the contingency should it arise, such as serious health risks—in some countries primarily through private insurance, in others through privately and publicly funded social security.

Individual or family, state or other (especially employer)

Saving up for retirement, either privately or through public programmes, can prevent poverty in old age. Reactive protection

Someone steps in to offset the effects of looming or actual emergencies— e.g. anti-poverty targeting programmes provide incomes, food or other necessities to poor families.

State or other than the individual or family in question

Individuals or families reduce their vulnerability in various ways, including: income diversification; drawing down of assets or borrowing when possible; adding family members to the labour force to raise more income; migration and remittances; and, in extreme cases, families and communities can sometimes rely on wider social networks. Actions such as distress sales of productive assets and keeping children out of school to work generate their own risks and/or future costs (Perry et al., 2007: 191). Taken together, these responses still fall far short of a good level of social protection. This is especially so when the share of families in distress rises quickly, as under conditions of macroeconomic recession, regional agricultural crisis or disease epidemic.5 When assistance comes substantially or wholly from the state, it is traditionally referred to as social protection. When the objective is straightforward, reactive poverty alleviation, it is referred to as ‘social assistance’ (e.g. welfare programmes), and when it is designed to prevent poverty or other emergencies, it is referred to as ‘social insurance’.6 Social protection is only one of a number of ways in which state actions affect the degree and frequency with which people suffer from poverty and other emergencies. People can be protected from temporary poverty or an income-related vulnerability by: • social assistance (social welfare); • social insurance, organized and at least partly funded by the state;7 • insurance mandated by the state but paid for partially or wholly by the individual and/ or the employer; • insurance organized and funded voluntarily by the individual;

Social protection, growth and employment • programmes with a social protection component (reflected in a subsidy) that also raise the primary income of poor people, such as microfinance; • any other programmes or policies that are pro-poor in terms of their impact on the primary distribution of income; and • any other programmes or policies that promote growth and, though not pro-poor (so inequality increases), do lead to a fall in poverty.

All of the above approaches can reduce temporary poverty and/or vulnerability. All except for insurance organized by the individual can reduce permanent poverty, since people who are permanently poor do not generally have the resources to self-insure. When the state is involved in social protection as usually defined (the first three items above), it may be the direct provider of protection or may mandate such provision through someone else, typically the employer (for more detail, see below). It may also subsidize provision through private suppliers of services, as with health. Social protection is sometimes thought of only in the narrow sense of redistributive policies8 such as welfare payments, but this is far too narrow an approach to guide thinking about overall strategies to minimize poverty and vulnerability. As noted above, individuals and families also respond to vulnerability through their own actions, when they can, and policy should aim to support and facilitate such actions, as by widening access to savings institutions.9 Also, there is no clear line between the policies that determine income levels and those that contribute to social protection. Employment and social protection issues need to be analysed simultaneously because they are causally linked to each other in many ways. Although their relative importance naturally varies widely across countries, the main forms of social protection include:10 • social assistance against poverty, taking such forms as welfare payments, food subsidies, cash transfer programmes and guaranteed employment programmes; • social assistance against specific threats to welfare, such as health problems (e.g. via a rural public primary health system); less frequently, protection against various forms of abuse (e.g. of children); • free provision of education up to a certain level for all children (which constitutes part of social protection for those in low-income families) and training programmes for low-income people; • a set of labour market policies, of which key elements are a minimum wage (or a set of minima attached to different types of job), limits on and financial recompense when the worker is fired, codes governing the hours and conditions of work and, less frequently, unemployment insurance programmes; • stabilization of food prices to avoid their becoming excessively high; and • at least some countries now also include microfinance and other programmes that help raise the incomes of small producers.

A conceptual framework

Coverage — the danger that poor people may be squeezed out of social protection programmes: Some social protection programmes are ‘universal’ in that they are in principle applied to the whole population, while others are deliberately targeted to, or by the nature of their organization apply only to, sub-groups of the population. Old-age pensions in industrial countries are virtually universal. Their main objective is often to have a social protection impact on the lower-income members of the society. This could also be achieved by targeting the pensions only at those lower-income people, but universality is preferred because it (i) avoids any possible stigma for those who receive the pension; (ii) avoids the risk that poor people will be missed through faulty targeting; and (iii) avoids the added costs of targeting, while (iv) being able to rely on an income tax system which assures that the bulk of the pension revenue accruing to well-off people will revert back to the state in the form of higher income tax payments. In principle, programmes that target poor people have the potential to achieve greater antipoverty efficiency —more poverty reduction per dollar spent—than other programmes. This does depend, however, on the costs and efficiency of the targeting process and also on stigma effects. One of the most effective ways to keep targeting efficient and economical is to design a ‘self-targeting’ system in which only those in the greatest need seek the assistance. Guaranteed employment programmes can achieve this by keeping the wage low enough so that only those in poverty will find it attractive. If the wage is relatively high, there will be too many non-poor people who want to participate in the programme, and cronyism and favouritism will then select in many of them, while selecting out many of the poor people. When self-targeting is not an available approach, other ways must be found to make sure that the resources reach poor people.

The broad mechanisms and institutions of social protection provision Each component of direct social protection (the first three categories listed above) involves a decision process, financing and delivery. The three main institutional participants are the state, the employer and the individual. The state’s role can be as mandator, as funder, as direct provider or in some combination of these roles. The other categories involve indirect provision of protection, since the benefits to the recipient come through some other channel, such as higher earnings of a microentrepreneur, the greater demand for labour due to a well-designed industrial policy, etc. In many settings, several social protection needs are bundled together, and their provision is part of the labour contract. Where social security provision is tied to formal employment and the administration lodged with the employer, the government is not the direct provider of most of the benefits involved, and the firm is typically acting on instructions from the state, which sets rules for how services are provided and often contributes to the fund from which they are financed. The content of such contracts is also affected by bargaining between employers and unions—for example, to top up the state’s minimum stipulations. Almost all (perhaps all) countries have some combination of employer and state involvement in social protection tied to employment, but the proportions vary.

Social protection, growth and employment

In many developing countries the employer plays a large role, with the insurance-related aspects of social protection thus limited to workers in the formal sector (large firms and the public sector) and excluding many in the urban informal and rural agricultural sectors. This strategy creates a dichotomy between those within the protective system and those outside it. One of the major issues arising in discussions of formalization policies is this inequitable social protection system. Recent trends towards making some elements of social protection universal have reduced that inequality but have created complex issues of how the two different systems may best co-exist and how to move toward a more universal system. Sorting out the various impacts of mandated employer-managed social protection arrangements is difficult. Taken at face value, the future social protection services received by the workers are financed partly by them through their withheld contributions to the social protection fund, with another contribution coming from the employer and sometimes a third from the state. But the real effects of such a system depend on how the worker and employment contributions affect the wage rate and on what these two types of agents would have done with the funds had they had unrestricted access to them. One possibility (or probability) is that the wage rate would be lower because of the non-wage labour costs the firm must pay. Normally, however, wages are not pushed down by the full amount of non-wage costs,11 so total labour costs to the firms are raised. Assuming that workers would not have saved the full amount of the additional wages they would have received in the absence of the withholding system, this means that they are engaging in some 足forced 足savings,12 with the state essentially forcing workers into a behaviour pattern which will either help them when services are needed (e.g. a health emergency) or in old age or will negate the need for the state to undergo expenditures at such times. The employer-based social security system typically becomes a combination of mandated minimum conditions for all covered workers, together with another set of usually less essential additional benefits. In Mexico, for example, there are eight components of the benefits package: health insurance, retirement pensions, disability insurance, profession risk insurance, life insurance, child care centres for workers, sports and cultural facilities and housing credits (Perry et al., 2007: 194). Some of these covered formal sector workers will naturally value certain of these benefits much more than others. Where an available benefit has no value to a worker, he/ she may see its implicit cost (wages forgone) as a pure tax. Most of these workers are not under the poverty line and might be in a position to handle more of their insurance needs on their own, either through private purchase or another of the risk-deterring steps available. The state is by definition involved in both the mandated employment-related contributory social protection systems and the non-contributory ones, so policy attention is naturally directed towards these two. Usually less attention goes to individual and family self-protection. But sometimes public policy can substantially influence the effectiveness of such protection, to the direct benefit of the people involved but also of the state, which thereby relieves itself of some part of the task it would otherwise face. Improvements to the financial system, such as the extension of rural banks with their deposit-taking functions in Indonesia and India at various points over the last few decades, appear to have made a significant difference (Patten and Rosengard, 1991; Burgess and Pande, 2005). Chile encouraged the development of private entities towards

A conceptual framework

which individuals could direct their pension fund. Many policies less directly related to savings can also play important rolesâ&#x20AC;&#x201D;for example, food price stabilization, which makes it easier for families to plan for future needs and against future emergencies.

Self-protection and the coverage of social protection systems Although world poverty trends have been moving in the right direction, at least until the current recession, in many individual countries progress has been absent or discouragingly slow; in some, trends have even been negative. Frequency and severity of vulnerability to emergencies are harder to measure than, for example, poverty or health. A rough idea of how successfully families deal with income shocks comes from analyses of the degree of consumption reduction per dollar of income reduction. Evidence from several Latin American countries suggests that on average a 10 percent income shock translates into a 2.5 percent to 4 percent decrease in consumption. Poor people are less capable of such consumption smoothing than others. In rural China, a 10 percent income shock lowered consumption by 1 percent for the wealthiest households but by 4 percent for the poorest (Jalan and Ravallion, 1999). Families are particularly vulnerable to the effects of health shocks (Baeza and Packard, 2006), which often drive them into poverty. One aggregate indicator of the possible performance of social protection programmes is the share of gross domestic product (GDP) spent on social protection. In the Latin America and Caribbean (LAC) region in the early 2000s this share ranged from almost zero (less than 0.5 percent in Jamaica and Ecuador) to over 16 percent (Uruguay). The regional average rose from 4.0 percent of GDP in the mid-1990s to 4.8 percent a decade later, compared to an average of 14.4 percent in Organisation for Economic Co-operation and Development (OECD) countries, 16.6 percent in Continental Europe and 8.3 percent in the USA (Perry et al, 2006, 189). When data are disaggregated, one sees that the bulk (usually over two thirds) is for social insurance, typically employer-managed. In a number of countries, however, its share has fallen as that going to poverty-oriented assistance programmes has increased. Transfers per person in the latter tend to be considerably smaller than in the traditional social security systems (Lindert et al., 2006). The ideal indicator of effortâ&#x20AC;&#x201D;still not often availableâ&#x20AC;&#x201D;is the percentage of GDP going to social protection services for lower-income or vulnerable people. Information on share of families covered by social protection programmes around the developing world remains inconsistent. Within LAC, where the data are more readily available, the share of the economically active population who contribute to pension systems varies from around 10 percent in Bolivia to over 60 percent in Costa Rica, with the regional average around 40 percent (Perry et al., 2007: 184). Despite a clear positive cross-country association with per capita income level, the ratios generally fell in Latin America between the 1990s and 2000s despite modest increases in per capita income. The reasons for this include increases in participation rates, especially for women during the 1990s, growth in the (usually lower-coverage)

Social protection, growth and employment

non-tradables sectors due to currency appreciation in the early 1990s (Colombia and Brazil), and social security reforms which may have lowered the perceived benefits of affiliation. Coverage rates of contributory systems are, as expected, consistently less for lower-income workers. No LAC country exceeds 50 percent coverage for the lowest quintile. In most, the poorest people are practically excluded (ibid.: 185â&#x20AC;&#x201C;6). In Brazil the coverage (percentage contributing) for the bottom quintile is a little under 20 percent and for the top one close to 70 percent. Inequalities in access tended to increase between the 1990s and the 2000s. In the Southern Cone countries (Chile, Argentina and Uruguay) and in Costa Rica this reflected falls in coverage of the poorest quintile. A number did launch or expand poverty reduction and/or social assistance programmes for poor people, so the situation is not as bad as these figures suggest. Still, the overall social implications of the observed trends call for a careful analysis. Pension coverage (of some sort) among elderly people ranges from around 10 percent to 80 percent (Uruguay and Brazil). For the few countries with separate data on contributory and non-contributory old-age pension systems, the latter usually accounts for a substantial share of the total coverage, and in Bolivia the huge increase in total coverage between the 1990s and 2000s is due to the broad coverage of the non-contributory system (ibid.: 188). A key component of social protection systems, both in terms of its importance to people and of its cost, is health care. Private insurance markets are of little help in this area, since the severe information problems that characterize both health issues and the provision of care are notorious for making the private market for risk prone to failure. Negative Âexternalities13 from a personâ&#x20AC;&#x2122;s not caring for himself/herself are another significant reason for public-sector involvement. Many innovations in subsidized public health care and provision of security against emergencies have been made recently. Some of these programmes reach mainly poor people, as with Oportunidades14 and Seguro Popular in Mexico. HIV and AIDS programmes have been successful in several African countries (Birn et al., 2009). But available figures continue to portray both low overall coverage of many health services in some countries and large income-related discrepancies of access in others. In Niger, for example, the share of births attended by medically trained personnel were under 10 percent for the bottom three quintiles, rising to over 60 percent for the top one (WHO, 2008: 28). While it is always hard to know what share of health situations require professional attention, an idea of the sensitivity of access to costs comes from the tripling of outpatient attendance in a region of Uganda which followed within a little over a year of the abolition of user fees (ibid.: 27). The increase came disproportionately from poor people.15

A conceptual framework

The dangers of classifying policies by main objective and disregarding interrelationships among them Although specific policies are usually directed mainly at a single societal goal, such as growth or poverty reduction, they usually also have other effects, often indirect ones. Partly to achieve simplicity and order in the policy-making and implementation processes, governments and their executing agencies tend to focus solely on the primary objective of each given policy. Up to a point, such ‘decentralization’ of policy-making and implementation is necessary, but it also brings dangers. Governments usually undertake joint decision-making of some form or other; however, there are many bureaucratic, human resource and other limitations and obstacles to effective coordination, and usually less joint decision-making occurs than would be desirable. Because of the rather different nature of the objectives, there is a particularly marked tendency to treat economic policies and social policies separately, and not to pay sufficient attention to making them consistent with and mutually supportive of each other. One might thus distinguish three broad orientations among economic and social policies: • exclusively or nearly exclusively improving market outcomes (growth, employment, equality etc.); • exclusively or nearly exclusively providing social protection (e.g. welfare policies for people clearly unable to earn a living on their own, health services for people whose economic performance is unlikely to be much affected by their health); and • significantly involving both market outcomes and social protection. Most people in decision-making positions place nearly all policies in one of the first two categories. In fact, although some ‘economic’ policies may legitimately fall into the first category and many ‘social’ policies into the second, many policies belong in the in-between range; although they have significant effects directly on market outcomes, they also constitute part of the social safety net. Sometimes, when it is feared that programmes with a social protection objective may create perverse economic incentives, the benefits provided are made conditional, as in conditional cash transfer systems or in cases where receipt of unemployment insurance payments is made conditional on some sort of training (as traditionally in Sweden). Such programmes thus provide conditional social protection and reflect consideration of both the social and the economic effects. The above ‘orientations’ reflect where the expected effects lie. But the actual effects may be substantially different: first, because policies may not be well designed or implemented vis-à-vis their intended objectives and, second, because they may give rise to many unforeseen indirect effects. One type of design shortcoming is the failure to take into account the interactions among different policies. The ideal in policy-making is thus to consider the impact of each

Social protection, growth and employment

possible combination of policies on each of the targets of policy and to select the recipe that can provide the best combination of outcomes. This involves distinguishing policies according to their combination of growth and employment/equality impacts on the primary distribution of income, and according to their links to the tax and transfer policies that affect the secondary distribution of income.

How a country’s economic structure affects social protection policy A country’s economic structure affects its social protection system in several key ways. The nature and severity of the needs for and potential supply of protection reflect the per capita income, its distribution and the dominant sorts of economic activities. • Higher average incomes allow people to take care of more problems on their own. Lower inequality means that the share who are poor or very poor and, therefore, most in need of protection is smaller. • Per capita income—which substantially determines the level of financial and human resources available for social protection—and the competence and honesty of the public sector determine the potential ‘supply’ of such protection. • The sectoral composition of economic activities affects the extent and nature of the needs for protection, because those needs vary due to factors such as seasonal smoothness of the income stream, frequency of severe drops in income etc. In low-income countries most of the population is in agriculture, which often implies severe fluctuations in income over the course of the year, as well as periodic crises through lack of rainfall or other misfortunes. • The dominant form of business enterprise has implications for how social protection services may best be provided and also for the availability of resources to provide them. Here the important dichotomy is between the larger formal-sector firms and the informal ones. As discussed below, the existence of a large informal sector creates a range of challenges for various elements of social protection policy. Since so many of the factors associated with social protection are associated with a country’s per capita income or level of development, it is no surprise that the systems differ markedly between countries at different income levels. Among major structural differences between lower- and higher-income countries are: • the declining prevalence of agriculture and/or rural residence; • the increasing size, role and capacity of the state; • the shift from the extended family to the nuclear family; and

A conceptual framework

• the increasing prevalence of medium-sized and large firms and of paid employment as opposed to self-employment. The major items in the social protection package are thus different between low-income countries — where the main challenges are famine relief; counteracting income fluctuations, chronic income poverty and disability; dealing with simple health issues; and assuring universal education at least to a basic level, all in ways as supportive as possible of the motors of pro-poor growth (usually small-scale agriculture and micro-, small- and medium-scale nonagriculture)—and high-income countries, where the main challenges are open unemployment, complicated and costly health issues, income poverty and disability, including age-related and single-parent-related poverty.

The key interactions between employment issues and social protection policy Social protection policy interfaces in a variety of ways with the more ‘economic’ policies in a country. Each of the elements of social policy aims to alleviate poverty or other social problems. But some have negative disincentive effects on such desirable behaviour as searching for a job, saving for retirement or engaging in entrepreneurial activity. The task for overall socio-economic policy is to find the best balance between positive support for human welfare and limiting any negative side effects. It is especially vital that social protection does not interfere with and, if possible, supports the sectoral motors of pro-poor growth, each of which is likely to create a good deal of employment, the focus of special attention here. The direct interface of social protection with mainly economic policies has three main dimensions: • incentive effects, where welfare or unemployment payments discourage people from searching for or keeping jobs; • investment effects, where cash transfer systems encourage parents to keep their children in school and thus raise the human capital stock of the country and lower future inequality; and • fiscal costs, since the programmes cost money in the form of direct fiscal outlays and related administrative costs, and there are costs to raising the tax revenues that are used in the programmes. It is important, however, to distinguish between direct and indirect interactions between social protection and employment outcomes. In many countries, the more important interactions are likely to be the indirect ones that link social protection policy positively or negatively to the motors of pro-poor growth. Among the indirect effects is any discouragement of economic activity connected with the collection of the taxes that fund the social protection expenditures.

Social protection, growth and employment

Positive links to the motors of pro-poor growth Given the central role of small agriculture and micro, small and medium-sized enterprises (MSMEs) in pro-poor growth, it is essential that social protection policy be as supportive of them as possible. In many developing countries, the main hope for equitable growth in the short and medium term is the small-scale agricultural sector. Making the sector the engine of income growth for those engaged in it, and having positive spin-offs on related rural and non-rural economic activities, entails the same broad requirements as the economy as a whole: technological improvement to raise productivity, investment, and the full and effective use of resources. The appropriate investment will include some in public infrastructure (irrigation works, roads etc.), some in private forms of capital (e.g. machinery, buildings) and some in human capital (education and training). Such investments would normally come from the traditionally responsible agencies or groups, but rural employment guarantee programmes (such as the National Rural Employment Guarantee Act in India) may also generate infrastructure. Making non-agricultural microenterprises more productive and profitable and advancing the performance of SMEs are the other general directions most likely to improve employment outcomes in the economy. Along the lines pioneered among others by the Grameen Bank in Bangladesh, microfinance may be provided along with various complementary services such as skills training. The potential contribution of large firms to pro-poor growth is greatest when they are labour-intensive (e.g. most clothing exporters), pay a large sum in taxes and/or have numerous links with more labour-intensive MSMEs. But, taken as a group, large firms are not labourintensive (Berry and Mazumdar, 1991), so, in the absence of positive linkages with the rest of the economy, they can pose a serious challenge to the achievement of pro-poor growth. One alleged reason for their failure to create more jobs is the set of labour regulations they face (see below). Subsidized unemployment insurance programmes should strive not to simply encourage more job-seeking in the modern large-scale or public sectors but rather to make the labour market function more efficiently for the employment-creating sectors that do hire a lot of workers, such as SMEs. By coincidence, many programmes directed to poor people will naturally target those working in these two pro-poor growth sectors, since they are the locale of most poor people in most countries.

Social protection policies with significant links to employment The main forms of social protection with potentially close ties to employment are: â&#x20AC;˘ guaranteed work programmes; â&#x20AC;˘ other work programmes designed for crisis situations, either of the economy or of the individual/family; â&#x20AC;˘ unemployment insurance, in its various versions, recently initiated in some least developed countries (LDCs);

A conceptual framework

• unconditional income supports, including pensions, triggered by age or retirement; • unconditional income supports triggered by poverty and/or inability to work; • unconditional food subsidies and other subsidies, which are targeted forms of raising real income; • conditional income support—for example, to poor families whose children (of a given age range) stay in school; • support targeted to specific needs, especially health and education, which can have major impacts on the supply of labour; • the minimum wage, firing rules and other employment regulations; • stance on the role of unions; • assistance to microfinance systems; and • assistance to NGOs, local groups or others involved in strengthening small-farm agriculture. Social protection policies can have a variety of different links to employment, some reflected in the demand for labour, and some in the level and skills available on the supply side. Table 1.2 illustrates part of the range of possibilities for selected components of social protection. Many effects, whether the direct ones that define the policy, indirect side effects or direct or indirect effects of conditionalities imposed, will vary with the details of the programme. Therefore, no general statements about links to employment can be based solely on theoretical analyses of programmes.

Table 1.2: Links between Social Protection Components and Employment, and Relations to Motors of Pro-poor Growth Panel A: Employment of Last Resort Demand For Labour

Supply For Labour

Basic mechanism

Direct job creation.

Possible side effects

Possibly substantial indirect increase in demand for labour through multiplier effects of the spending by beneficiaries.

Possible conditionality effects

If conditional on being openly unemployed, may discourage formal-sector employment.

Possible links to motors of growth

Produce infrastructure for agriculture and thus raise agricultural production.

Basic mechanism

No presumed effect.

Possible side effects

Negative side effect on supply to other sectors, especially if conditional on being unemployed.

Possible conditionality effects

If conditional on being unemployed, may discourage job search; potential training/learning effects.

Possible links to motors of growth

Training may raise agricultural productivity.

Social protection, growth and employment

Panel B: Food Price Stabilization Demand For Labour

Supply For Labour

Basic mechanism

No presumed effect.

Possible side effects

Possible indirect increase in demand for labour if stabilization makes small agriculture more profitable and less risky; effects will differ according to whether the country is usually a food importer or exporter.

Possible conditionality effects

None presumed.

Possible links to motors of growth

See â&#x20AC;&#x2DC;possible side effectsâ&#x20AC;&#x2122;.

Basic mechanism

No presumed effect.

Possible side effects

Any effect is of unpredictable direction; if supply is reduced, this will likely be desirable (as where an increase was necessitated by poverty brought on by high food prices).

Possible conditionality effects

None presumed.

Possible links to motors of growth

Should raise long-term attractiveness of small-scale agriculture unless the country is a significant food importer.

Panel C: HIV/AIDS Policy Demand For Labour

Supply For Labour

Basic mechanism

Skilled personnel plus some others.

Possible side effects

Better health will raise output and the demand for labour.

Possible conditionality effects

None presumed.

Possible links to motors of growth

Much of the effect on output will be in the two propoor growth sectors.

Basic mechanism

Raise supply by saving lives and improving health and productivity of people living with HIV/AIDS.

Possible side effects

Raise supply of labour to the market by freeing up family members otherwise engaged as caregivers.

Possible conditionality effects

No presumed effects.

Possible links to motors of growth

Increased supply and productivity of workers will mainly impact output in the two key sectors.

Issues/distinctions in the design of social protection policy Consideration of the benefits and costs of any given form of social protection requires that distinctions be made by context (since all relationships vary at least to some extent by setting) and also by the set of people (target group) likely to need social protection. Target groups are

A conceptual framework

defined by age, with special attention to children and older people, and by ability/inability to work and earn a living, where the main subcategories are people with disabilities or those that are physically ill, and single parents who must care for their children. Some forms of social protection are by definition targeted at special groups; others are not, but their logic and potential payoff varies by group. The social protection requirements needed to prevent social crises are affected greatly by cultural norms, since in traditional societies all of these groups would, to a greater or lesser degree, be taken care of by relatives, friends, community etc., but these modes of protection tend to erode as societies become more industrialized. To summarize, a taxonomy of social protection policies and issues needs to distinguish: • the type of support (work guarantee, health etc.) and the likely extent of the benefits; • to whom provided — for example, specific groups (pensions to older people, income support to single parents) as opposed to universal; • by whom provided, i.e. the state or the employer (if the individual handles his/her own arrangements in full, there is no social protection as such); • the design of the social protection provision system — for example, unemployment insurance generously provided for long periods vs. small payments provided for short periods and under stringent conditions. At the interface between social protection and alternative types of support, there is often an argument for keeping social protection less than generous so as not to act as a disincentive to other forms of help which may be cheaper and/or more efficient; and • the possible positive and negative side effects, related among other things to the alternative ways of confronting the poverty or emergency. Although the specific instruments involved in social protection are numerous, almost all of them fall under a few categories of broad objectives, which may be usefully classified as: • increased income/purchasing power per se (e.g. social assistance in the form of income payments); • increased employment leading to increased income, as with guaranteed employment programmes and many other instruments that affect employment; • increased income and/or employment security; • improved health; • increased health security; and • improved education. Many instruments falling under a given category naturally tend to have somewhat the same side effects, positive and negative. For example, social assistance in the form of income payments often has the potential to discourage the search for work: at one end of the spectrum of beneficiaries are those who cannot work;16 at the other end are those who are able to work but prefer not to. Unemployment insurance may also discourage the search for work or encourage

Social protection, growth and employment

temporary types of work (with the connivance of the employers when generous terms allow it), as well as costing money. On the plus side, any instrument that increases income may discourage illicit activities and/or have positive multiplier effects. Instruments that foster employment are expected to provide an output of some value and hence to be less costly, net, than unconditional income payments. They may provide worker training and avoid the scarring effects of unemployment 17 (that curtail the learning of both skills and pro-work attitudes that go with employment), but they may be hard to organize effectively. Although the health and educational components of social protection are thought of as raising welfare directly, the positive economic effects of the resulting greater level of labour productivity can be seen as side effects. Other examples of the range of policy considerations for specific components of the social protection package (the minimum wage, non-contributory pensions for low-income retired workers, guaranteed work programmes and unemployment insurance) are given in Table 1.3.

Table 1.3: Key Considerations around Selected Social Protection Policies Panel A: The Minimum Wage Main objective

Assure that wages achieve a reasonable minimum level and thereby place a floor under the incomes of lower-income families.

Possible positive side effects

May raise wages in the informal sector as well as the formal sector to which it is applied.

Possible negative side effects

May discourage hiring in the modern formal sector (where the regulation is implemented) and thereby increase poverty and labour inequality.

Ambiguitya about net effects

High — it is empirically difficult to estimate most of them with precision.

Degree of debateb about effects

High, due among other things to rigidly held views of unions and defenders of labour, businesses and some economists.

Sensitivity of net effects to design and implementation

High. Where the minimum wage is low, it is highly likely that its impact on workers will be on balance positive, but of modest proportions. When it is very high, it is possible that the net impact can be strongly negative, but this depends on elasticities of the demand for labour. When it is intermediate, the effect is also hard to predict.

Direct costs

Low — just the cost of monitoring.

Main substitute programmes/ policies/conditions

Strong overall demand for lower-skill, lower-wage labour; guaranteed employment programmes; strong growth of the SME sector etc.

Main complementary programmes/policies a ‘Ambiguity’ refers to the true difficulty in assessing impacts. b ‘Degree of debate’ reflects some combination of true ambiguity and the presence in the debate of vested interests and other groups who do not rely strongly on professional judgements of the effects in question.

A conceptual framework

Panel B: Non-Contributory Pensions for Low-Income Retired Workers Main objective

Raise incomes of lower-income elderly families.

Possible positive side effects

May generate positive multiplier effects in the local economy if such people are numerous and there are underutilized resources that can produce items the elderly people will buy.

Possible negative side effects

It is unlikely that the programme would discourage work activity or savings on the part of the first set of beneficiaries, but such effects may arise among working adults who now predict that the state will provide such pensions when they reach retirement. For very low-income people with great difficulty in saving, this effect is unlikely to be significant. For somewhat better-off people, it might be; this is an empirical question that must be studied.

Ambiguity about net effects

Low, since the benefits of modest pensions to low-income people are presumed to be very high per dollar.

Degree of debate about effects

Low, as per the above.

Sensitivity of net effects to design and implementation

Moderate with respect to the first group of beneficiaries, unless design and implementation are quite badâ&#x20AC;&#x201D;e.g. the funds do not reach those for whom they are intended. Higher with respect to later groups, due to lack of study to date.

Direct costs

Varies with the age group and income range covered, but can be substantialâ&#x20AC;&#x201D; e.g. around 1 percent of GDP, as in Brazil.

Main substitute programmes/ policies/conditions Main complementary programmes/policies

Panel C: Guaranteed Work Programmes (Employment of Last Resort) Main objective

Assure that workers achieve a minimum level of labour income when employment is especially difficult to obtain; produce useful output, such as infrastructure.

Possible positive side effects

Can help to provide a wage floor in the relevant labour market.

Possible negative side effects

May discourage search for jobs elsewhere.

Ambiguity about net effects

Moderately high. The value in providing an income floor is relatively well accepted unless implementation is weak. Value of resulting product appears to vary rather widely.

Degree of debate about effects

Modest.

Sensitivity of net effects to design and implementation

Moderately high, as per the above.

Direct costs

Low, just the cost of monitoring.

Main substitute programmes/ policies/conditions

Income supplements; unemployment insurance.

Main complementary programmes/policies

Well-designed investment programmes to which these programmes can be appended.

Social protection, growth and employment

Panel D: Unemployment Insurance Main objective

Assure that wages achieve a reasonable minimum level and thereby raise incomes of lower-income families.

Possible positive side effects

May raise wages in the informal sector as well as the formal sector to which it is applied.

Possible negative side effects

Discourage hiring in the modern formal sector; increase poverty and labour inequality.

Degree of debate about effects

High.

Ambiguity about net effects

High.

Sensitivity of net effects to design and implementation

Direct costs

Low, just the cost of monitoring.

Main substitute programmes/ policies/conditions

Strong overall demand for lower-skill, lower-wage labour; guaranteed employment programme; strong growth of the SME sector etc.

Main complementary programmes/ policies

The shift towards two-tier social protection systems A number of developing countries have introduced non-contributory social protection systems or partial systems in the last decade or so. These are essential because of the poverty levels of those outside the previous social protection systems based on the formal sector. Because the informal sector houses a high share of the poverty in most countries, such non-contributory systems cannot be linked to formal-sector employment, so alternative delivery mechanisms must be found, which makes the social protection system more complicated. In pursuit of an effective social protection system special attention should be directed to those groups in the grey area between the definitely formal and the definitely informal components of the labour market and the corresponding segments of the social protection system. These people may wind up in the formal sector or the informal one depending on a variety of factors. Within this group there is an important distinction between those who are voluntarily informal and those who are involuntarily so. In the absence of any universal social protection, the two modes of provision are through employment-linked social security and through private initiative. The benefits of such a system are that:

A conceptual framework

• those covered receive social protection; the social importance of the benefits depends on where they are in the economic hierarchy of the country and their potential for self-protection; • the system achieves a degree of self-financing, thereby partially relieving the state of the responsibility; • it may raise aggregate savings through forced savings of workers and perhaps of firms; and • administration may be more efficient than if the full responsibility lay with the state. The problems with this system in a low-income country are that: • many poor people are left out, since they cannot purchase private insurance or keep themselves out of poverty; • many systems do not allow for portability of one’s social protection privileges across different jobs and/or require a minimum number of years of contribution for access to some benefits. In countries with large informal sectors, many workers move in and out of the formal sector with some frequency and become disqualified for benefits (Perry et al., 2007: 196).18 Such systems introduce horizontal inequity among workers as well as inefficiency, by discouraging possibly desirable mobility;19 • in some cases all working family members must contribute even though other family members would be covered by one’s own package. The contribution of the second worker becomes a pure tax.20 Benefits need to be more tightly linked to contributions; • the system may be inefficient in other ways, such as where administrative costs are high, the wrong services are provided etc.; • the upward pressure on wage costs may discourage employment creation in the formal sector and worsen overall income inequality; and • it may discourage formalization, which may bring growth benefits, an increase in decent employment etc. This effect can come through the disincentive to firms hiring because of the high labour costs and rigidities and/or through opting out by persons dissatisfied with the social protection package available. These various potential benefits, costs and risks are likely to vary widely by setting and by group of workers, so their analysis requires knowledge of behaviour patterns of the relevant groups, and how different social protection systems can affect each one. The health component of social protection systems is especially important, since health is an area where protection is extremely valuable to people. How it is handled in the various systems can, therefore, play an important role in the overall efficiency of the combined system, in whether workers are excessively drawn to one or another sector by differences in the health protection available. Health exemplifies the complexities that can arise in the two-tier systems now becoming more common.

Social protection, growth and employment

In this area there is a strong case for public involvement, since health crises are so cataclysmic for many people, and because of the ineffectiveness of private markets, especially for those who might be just barely able to pay. Given the negative externalities of sickness, there is a clear risk management rationale to take the choice of cover out of the hands of employers and workers. But public health services have often not conformed to sound insurance principles and have often not delivered on their promise of quality services (ibid.: 200). Where outreach has been widened, this has led to low quality and doubtful fiscal sustainability. In recent years governments have opted for more sustainable, minimum-benefit types of packages. Any additional coverage beyond minimum should be on a contributory basis. In this area constitutional rights to health have contributed to defective systems, at least in the short term.

The implications of high levels of informality for interactions between employment and social protection Formality is not a simple concept, one of the reasons for the lack of a clear general answer as to whether it is something that should be encouraged by policy. Two main aspects of formality, as usually understood, are compliance with tax regulations and with labour market regulations. The societal desirability of a firmâ&#x20AC;&#x2122;s compliance with labour regulations depends on the quality of the regulatory apparatus and on the type of firm. Two related but different debates have arisen around the links between social protection and informality. First, to what degree does the presence of informality on a large scale make social protection policy more difficult and more costly? Second, when non-contributory social protection instruments are developed to deal with the problems of poor people independently of their attachment to formal-sector firms, does this further encourage informality, as Levy (2008) contends, and if so, does it come at the expense of economic efficiency and growth? It is widely accepted that a large informal sector can hamper the effective implementation of social protection (e.g. PagĂŠs, 2009); administration of many programmes is easier in a more formalized system, with good records on many relevant types of information, developed interactions between firms and the state, and so on. But the practical policy question is whether the difficulties in the social protection area that result from informality constitute a significant argument to contemplate pushing formalization as a policy, and if so, whether such a push could meet with enough success to make it worthwhile. There are several potential strands to the argument for a pro-formalization policy. One is that since informal firms do not pay taxes, their informality constricts the ability of the state to spend, whether on social protection or on anything else. Yet the taxes paid by a small, previously informal and now formalized firm are unlikely to be very high. The argument is much more persuasive when it includes the view that formalization of currently informal firms would open the way to advances in productivity and incomes, and thereby both promote economic growth and raise government revenues. But this view also remains open to considerable debate (Berry, 2010).

A conceptual framework

Probably the most persuasive component of the contention that the informal sector is a problem for social protection policy is that important parts of social protection are better administered by the employer than by the state. This is arguably the case with respect to larger firms, but the small and often unstable formal firms (to which formalization of small and unstable informal firms will most likely give rise) are hardly ideal for managing social protection services. The prevalence of informal firms itself implies a looseness in the labour market which is likely in any case to detract seriously from the general implementation of some standard components of the worker protection package, including rules on firing, hours of work etc. While it would certainly be simpler to be able to deal with some of these problems as they are dealt with in industrialized countries, there is little realistic chance of doing so in an economy dominated by the informal sector. What of the argument that the new forms of non-contributory social protection encourage informality? The possible disincentive to look for work/incentive to remain unemployed in developed countries, where unemployment insurance is the norm, has a sort of developing-country counterpart in the possible disincentive facing small informal firms to switch to formality or, conversely, the incentive for formal firms to become informal. Where combined costs to firm and workers of opting for formality (with its tax and labour regulations) exceed the private benefits, but that formality could have brought social benefits, private decisions will diverge from the social optimum. Thus, the discussion about desirability of formality includes the extent to which social protection benefits not attached to formal employment are seen as possibly constituting a disincentive to formalization of firms. Regardless of the pros and cons of a push towards formalization, the implications of informality and the growth of non-contributory and non-employment-related social protection provision add new twists to the already complex links between employment issues and social protection. Poverty makes non-contributory protection more important. Informality makes employment-linked social protection unfeasible for much of the population. The increasing fragmentation of social protection systems may create an additional disincentive to formalization.

Impact of the current world recession Even though the current crisis started in the industrialized world, developing countries have been affected through different channels, most notably by a fall in exports, a worsening of the terms of trade, a drop in financial flows and the decline in remittances from migrant workers abroad. These negative consequences have led to rising unemployment and underemployment and an overall reduction in income from labour activities. Recent signs that the world crisis may be technically coming to an end do not necessarily mean a recovery of employment any time soon. It is well known that job creation tends to lag behind output recovery, with lags often extending several years.

Social protection, growth and employment

The medium-term post-crisis years are also a source of concern. There are clear indications that the conditions that supported rapid growth in developing countries during the pre-crisis years will not be there any longer, or not to the extent they were before. One area of rising concern is that the space that developing countries had to increase exports to developed countries, and thus leverage growth, may be significantly smaller in the coming years. The post-crisis years face a world economy likely to grow more slowly and to be less tolerant of large currency imbalances and thus unable to absorb rapidly growing volumes of imports. Partly because of this, and in the absence of better policies, dynamic employment creation may follow only with a significant lag, even if considerable economic growth resumes. The need to innovate and reorganize production and service delivery to reduce costs might well mean a weaker demand for labour. Less rapid growth and lower employment intensity of growth would make inclusive growth and progress in human development more difficult.

Some general principles Although the optimal social protection package naturally varies by country, there are a number of general guidelines of value in the successful design of overall social protection policy and strategy. These include: 1. Other things being equal, the main components of a social protection system should aim to connect with and benefit poor people first and the less poor later. Sometimes the most effective way to do this is to have a universal programme; sometimes this is best achieved by targeting. Sometimes this group is best reached by indirect effects of one or another component of social protection. In the overall design, not all components have to reach the poorest people (or perhaps not reach them immediately), but this objective must be a high priority, since there are almost always political and administrative obstacles to achieving it. 2. As much as possible, social protection programmes should be linked positively to the motors of pro-poor growth. For example, rural employment guarantee programmes may generate infrastructure (irrigation works, roads etc.) of value to the small farmers of the region. Such linking will sometimes occur almost automatically, and sometimes it will require thoughtful innovation. Sometimes it will benefit from careful coordination with existing programmes, and sometimes it will be new, and the choice of the best institutional home for it will require some consideration. 3. Like good overall economy policy, good poverty reduction policy, of which social protection is an important component, should be as internally coherent as Âpossibleâ&#x20AC;&#x201D;i.e. its various components add up to more than the sum of the parts and do not work against each other but, rather, constitute a mutually supporting system. Thus, if the optimal type of development in a given rural/agricultural area involves raising agricultural productivity and facilitating the growth of non-agricultural activities by the same

A conceptual framework

families, the optimal social protection policy package is likely to include investments to raise agricultural productivity (including infrastructure, better crop varieties etc.) with labour-intensive techniques used whenever possible; microcredit to help small non-agricultural enterprises to grow; and some forms of training to raise entrepreneurial capacities. The more these programmes can be linked to each other through consistent timing and consistent messages, the better. Typically such links are weak or incipient, since different agencies are responsible for different components, and the potential advantages of integration are either not fully recognized or are institutionally hard to carry out. 4. Careful thought should be given to the best way to implement any given programme, including which institutions should be involved, whether it should be self-targeting and/or conditional, and which complementary services may be provided along with the core ones. Such care is essential because all programmes have costs and side effects (both positive and negative) which must be considered in the overall picture of costeffectiveness. Good programmes require great attention to detail; with worrisome frequency, failure to pay attention to the detail detracts greatly from overall effectiveness and can even imply that a programmeâ&#x20AC;&#x2122;s net effects are negative.21 Good design usually draws not only on the record of experience in the country but also on that of similar programmes in other countries; it also takes advantage of monitoring to assure that easily detectable problems are dealt with quickly, and periodic ex-post evaluations give a more in-depth picture of how well it is working. Making such checks workable requires advance thought to assure that the relevant information is collected as the programme proceeds. It is important to collect data on administrative costs as a ratio of funds transferred and on leakages or other misdirection of funds. 5. A special complication arises in those countries where substantial elements of social protection occur through the employer, but this system excludes the great majority of the population (if the informal sector is large) and may also have a variety of negative effects on employment generation. The situation calls for a shift in public provision to provide social protection to the poorest groups. This, however, leaves the country with a markedly dual system, which raises a number of complications, both in general and in terms of the interface with labour market outcomes. Since any attempts to deal with those complications are recent, it is important for both the countries involved and others to learn as much as possible as quickly as possible from them.

Social protection, growth and employment

Annex 1.1: Definition of some social protection terms A. Social protection Social protection is defined as the set of policies and programmes designed to reduce poverty and vulnerability by promoting efficient labour markets, diminishing people’s exposure to risks and enhancing their capacity to protect themselves against hazards and interruption/loss of income. The policies and procedures included in social protection involve five major kinds of activities: labour market policies and programmes, social insurance programmes, social assistance, micro and area-based schemes, and child protection. ‘Social safety net’ and ‘social security’ are sometimes used as alternative terms to ‘social protection’. The Asian Development Bank Interdepartmental Working Group on Social Protection found that in describing the range of concerns covered by this strategy the term ‘social protection’ was the more common in international usage and decided to use that term. ‘Social safety net’ appears to have a less precise meaning. Some people use it to mean the whole set of programmes and policies discussed in this strategy, while others use it to refer only to welfare and targeted programmes for poor people. On the other hand, ‘social security’ is used to refer to the comprehensive mechanisms and coverage in developed countries and is less applicable to new areas such as community, micro and area-based schemes.

B. Labour market Labour market policies and programmes are designed to facilitate labour adjustments and promote the efficient operation of labour markets. ‘Labour market’ is a term used in labour economics to indicate the exchange between the demand and supply of labour for a nation, region, industry or occupation. The outcomes of this process include the level of (un)employment, pay and employment conditions. Labour codes are regulatory frameworks containing provisions on labour relations as well as standards for working conditions including, for example, provisions on working hours, minimum age, minimum wage, occupational safety and health, social security and provisions regarding the special needs of women. Active labour market programmes are interventions to encourage people to take positive action to improve their employment prospects. They include vocational training and retraining, employment services, job search assistance and job fairs.

A conceptual framework

Labour exchanges and employment services are designed to help match employers’ job vacancies with workers looking for jobs. They may be operated by the state, particular employers or groups of employers, non-governmental organizations (NGOs) or the private sector.

C. Social insurance Social insurance programmes cushion the risks associated with unemployment, disability, work injury and old age. These programmes are usually tied to replace a specified percentage of a worker’s previous earnings, with the percentage a function of work years, level of pay and often social criteria such as family size and credit for certain non-work periods (e.g. child care and registered unemployment) or below-average earnings. These programmes may be contributory or non-contributory. The former are financed, at least in large part, through earnings-related contributions from employers and/or employees, whereas the latter are financed from the general budget. Eligibility for benefits under contributory programmes is usually limited to those who have made a minimum number of contributions. Demogrant is a non-contributory benefit paid to individuals solely on the basis of meeting specified demographic criteria (e.g. age, residence or disability); normally demogrant benefits are financed on a pay-as-you-go basis. Unemployment insurance is a programme designed to replace a percentage of prior earnings in covered employment for those who have lost their jobs. Benefits are usually limited to those who were previously employed and tend to last for only a specific period of time. Unemployment insurance is tied to job search requirements and registration with labour exchanges but is not means-tested. Unemployment insurance is usually designed to deal with frictional and structural unemployment. In its latter aspect, it is a tool of macroeconomic policy as well as a means of social protection in that it allows consumption levels to be maintained in the face of temporary declines in employment and output. Severance pay programmes mandate that employers provide workers who are terminating their employment with a sum of money related to past periods of employment and pay. They usually are a liability of the employer for which separate provision has to be made (see ‘funding’, below). Work injury or workmen compensation programmes are designed to compensate workers for work-related injuries or diseases. Employers pay premiums to finance costs, with the premiums often related to an employer’s claims experience. In some countries, these programmes, though mandated by law, are run through private insurers. Disability or invalidity insurance (pension) is usually linked with old-age and survivors’ insurance. Such programmes cover both non-work and work-related injuries or diseases that render a person incapable of working in full or in part. Some countries have programmes only for permanent and total disability. Others have programmes that compensate workers for disabilities that limit the hours or type of work they can perform. Although tied to initial and

Social protection, growth and employment

ongoing reviews of work capacity, disability insurance programmes are not means-tested. Most countries provide disability insurance through public programmes, usually financed on a payas-you-go basis, but some countries have moved to requiring the purchase of disability cover through private insurance companies. Survivors’ insurance (pension) is usually linked with old-age and disability insurance. Two types of protection are involved: (i) death of the insured before retirement, and (ii) death of a pensioner in retirement. In the case of ‘young survivors’ insurance, the primary focus is on surviving children and the surviving caregiver parent, although other dependent family members may be covered. In the case of post-retirement death, the focus is on the survivor of an aged spouse-pensioner. Most countries provide survivors’ protection through public programmes, usually financed on a pay-as-you-go basis, but some countries have moved to requiring the purchase of term insurance for young survivors through private insurance companies. Old-age insurance or pensions are usually linked with disability and survivors’ pensions, but the options for old-age provision are larger. • Pensions, strictly construed, are periodic payments after retirement for the life of the pensioner (and usually surviving aged spouse). Many countries’ old-age schemes, however, only allow for lump sum withdrawals upon retirement (e.g. provident funds—see below), leaving it to individuals to allocate the sum over their retirement years. • Defined benefit schemes are ones in which a worker’s promised pension or lump sum is a function of a formula tied to years worked, earnings and social criteria and in which some entity—state, employer, insurance company—bears the financial risk of the promise or entitlement embedded in the formula. Defined contribution schemes are ones in which a worker receives an accumulation based on contributions and investment returns and bears the financial risk. Many hybrid schemes exist that contain both defined contribution and defined benefit elements. • Many countries finance their mandatory old-age (and related disability and survivors’) pensions through a state scheme on a pay-as-you-go basis — i.e. current revenues, mostly from a specific contribution (payroll tax), are used to cover current benefits. (Such schemes, however, will often have some assets to cover periods of cyclical unemployment or negative wage growth). Some countries finance their mandatory monopoly state pension schemes through a combination of pay-as-you-go financing and funding (see below). These so-called partially funded schemes have large asset reserves designed to hold down future contribution levels. • Many countries finance old-age pension provision on a fully funded basis. This might be done through a state monopoly scheme, which is usually called a provident fund. Another means is a legal mandate imposed on employers, workers or both to accumulate assets in a special or general purpose financial intermediary (e.g. as in Latin America and Europe) equal to promised benefits or targeted account levels. Such mandated pensions are usually defined contribution pensions.

A conceptual framework

• Many countries also have voluntary occupational pension schemes that are taxadvantaged such that accumulations receive a tax-free return, and lifetime compensation is more lightly taxed. These tax advantages are sometimes extended to voluntary individual retirement savings. • Countries often support mandatory and, in some instances, even voluntary funded pensions with explicit guarantees, thus introducing a defined benefit element to what might otherwise be a defined contribution system. These guarantees create contingent liabilities for governments. • Some countries require coverage through both a universal state scheme (entirely payas-you-go or partially funded) and a fully funded scheme that operates as a mandate on employers or workers. The term ‘funding’ has several meanings. The strictest meaning is that the entity that is responsible for the promised pension or the retirement accumulation must hold title to a diversified portfolio of marketable assets equal to the promised pension or the recorded accumulation. However, the term is also used to cover many provident schemes where the accumulation is backed solely or primarily by government bonds that are not generally marketable. Similarly, the term is used by some to cover cases where an employer is making a pension promise backed solely or primarily by its future earnings (‘book reserve’). Universal programmes are not contribution-related and are financed from the general budget. In a few countries, such benefits—typically a children’s allowance or a citizens’ old-age pension—are a function only of age and citizenship (or residence), though they may be taxable.

D. Social assistance and welfare service Social assistance and welfare service programmes protect those with no other means of adequate support. A few countries use means-tested programmes to cover a very large proportion of the population against the contingencies of old age, disability, death or unemployment. Most countries, however, use a combination of social insurance programmes and means-tested social assistance, with the latter reserved for filling the gaps. Social assistance includes government and non-government institutions that provide residential care and in-home services to particular groups such as orphans and frail elderly people. Four main targeting methods are available: • Group targeting provides benefits to a specific population group (e.g. mothers, children or communities in a particular geographical area); this is the easiest targeting method, but leakages to the non-poor are normally large. • Means testing: means-tested targeted programmes provide benefits to households below a certain income level; means testing is more accurate in targeting poverty but has larger transaction costs than group targeting.

Social protection, growth and employment • Proxy means testing is a subset of means testing, in which targeting is done through other (easy-to-collect) indicators or proxies to correlate the level of income/poverty of beneficiaries; on a temporary basis, it is possible simply to give discretion to local public officials or to a local NGO. • Self-selection avoids moral hazards by imposing disincentives to participants in programmes, either because the benefits are too low or because there is some social stigma associated with them (e.g. food-for-work programmes and public works because normally only those truly in need accept them). However, given that benefits are normally below market wages, these programmes are not effective in bringing people out of poverty.

Most social assistance programmes are means-tested. Formal means testing involves conditioning benefits as a function of a person’s or household’s other income and assets. Some means tests exclude some income and particularly some assets such as the beneficiary’s house and a minimum amount of agricultural land. Where formal means tests are not practical, informal approaches to targeting benefits, or proxy means testing, can be used. These involve using proxies such as household size and composition (e.g. single parent), geographic area, age and disability as an alternative measure of the need for assistance.

E. Micro and area-based schemes Social funds are financed by national governments to operate in local communities, sometimes through NGOs, to provide temporary employment in public works and services. Projects selected tend to be those that will improve the local social infrastructure, such as schools, hospitals and roads. The projects are usually designed to create new facilities and provide employment opportunities to unemployed and underemployed local residents. Agricultural insurance is a financial mechanism that aims to reduce the uncertainty of loss by pooling together large numbers of uncertainties so that the burden of loss can be distributed. Agricultural insurance extends the concept of insurance to the different types of agricultural undertakings. Crop insurance is a mechanism through which farmers are protected from losses due to natural causes, including flood, fire, drought, disease, pestilence, wind, hail etc. Crop insurance can cover a variety of agricultural pursuits, including grain cultivation, livestock, beekeeping and prawn cultivation. Crop insurance tends to be financed by a levy equal to a modest percentage (usually less than 5 percent) of the value of the crop being insured. Insurance payments may compensate most (e.g. 80 percent) to virtually all of the loss. Agricultural insurance has so far had very limited success, but it is an area of possible intervention if adequately accompanied by reinsurance mechanisms. Reinsurance for agricultural insurance allows carriers of agricultural insurance business access to the underwriting capacity and capital of the world market. However, such reinsurance facility can be provided only to technically viable agricultural insurance programmes. It is also important for international reinsurers to have trust and confidence in the objectivity of loss estimations, absence of adverse selection/moral hazards, and independence of the programme

A conceptual framework

from political interference. A reinsurance facility may be proportional or non-proportional. A proportional or quota share facility is one wherein the reinsurer takes a share in the premium collected (e.g. 40 percent) and assumes the same percentage of a loss. In a non-proportional facility (an example is an excess of loss cover), the reinsurer quotes a fee (sometimes unrelated to the premium charged by the insurer to its policy holders) against which it will pay a loss when it exceeds a set percentage of the sum insured up to a higher designated percentage of the sum insured. Reinsurance is an effective risk-transfer mechanism for covariant catastrophic risks to spread the burden of loss globally. It can be a major source of much-needed foreign exchange that can provide a stabilizing effect when a country is reeling from a huge loss. Microinsurance refers to voluntary and contributory schemes for the community, handling small-scale cash flows, by way of both income and expenditure, to address community risks. Such schemes are often of a local character and have a very small membership. The primary aim of many of these schemes is to help their members meet the unpredictable burden of outof-pocket expenses, such as a hospital emergency, death or funeral expenses. Microinsurance offers the possibility to provide social insurance at affordable prices, expand coverage by having a realistic understanding of the problems that communities face, and promote community involvement.

F. Moral hazard and adverse selection problems Moral hazard refers to the presence of incentives for individuals to act in ways that incur costs that they do not have to bear. Adverse selection may occur when the act of providing support against certain risk can increase the likelihood that the undesired risk occurs.

Table 1.A1: Asian Development Bankâ&#x20AC;&#x2122;s Components and Subcomponents of Social Protection Labor Market Programmes Direct employment generation (microenterprise development and public works) Labor exchanges and other employment services Skills development and training Labor legislation (including minimum age, wage levels, health, safety, etc.) Social Insurnace Programmes Programmes to cover the risks associated with unemployment, sickness, maternity, disability, industrial injury, and old age Source: Derived from I. Ortiz (ed.) Defining an Agenda for Poverty Reductionâ&#x20AC;&#x201D;Proceedings of the First Asia and Pacific Forum on Poverty, Volume 2, p. 57, ADB, Manila, 2002, ADB, Social Protection Strategy, pp. 14â&#x20AC;&#x201C;22, Manila, 2001.

Social protection, growth and employment

Social Assistance and Welfare Programmes Welfare and social services targeted at the sick, the indigent, orphans, and other vulnerable groups Cash-in-kind transfers (e.g., food stamps) Temporary subsidies for utilities and staple foods Micro and Area-Based Schemes Microinsurance schemes Agricultural insurance Social funds (usually involving the construction, operation, and maintenance of small-scale physical and social infrastructure) Disaster preparedness and management Child Protection Early child development activities— e.g., basic nutrition, preventative health, and educational programmes Educational assistance (e.g., school-feeding, scholarships, fee waivers) Health assitance (e.g., reduced fees for vulnerable groups) Street children initiatives Child rights and advocacy/awareness programmes against child abuse, child labor, etc. Youth programmes to reduce health risks (especially HIV/AIDS and drugs) and anti-social behaviour Family allowances (e.g., in-kind cash transfers to assist families with young children to meet part of basic needs) Source: Derived from I. Ortiz (ed.) Defining an Agenda for Poverty Reduction—Proceedings of the First Asia and Pacific Forum on Poverty, Volume 2, p. 57, ADB, Manila, 2002, ADB, Social Protection Strategy, pp. 14–22, Manila, 2001.

Notes 1 For similar definitions of the appropriate policy package see Khan (2001), Osmani (2004) and Zepeda (2004). 2 There have been debates regarding the definition of pro-poor growth. The definition presented here is in line with the thinking emanating from UNDP’s International Policy Centre for Inclusive Growth (IPC-IG, formerly known as International Poverty Centre) as articulated by Kakwani, Khandker and Son (2004). For some details on the debate see, for example Zepeda (2004). 3 It is important to note that the economic concept of ‘productivity’, related to economic efficiency, does not refer to labour productivity (output per worker or per labour hour) but to the ratio of the value of output to the amount of all inputs used, labour, land, capital etc. 4 Regardless of the precise definition of poverty used, it is important to make a distinction according to severity, since poverty is, after all, a matter of degree, and also between chronic and temporary poverty. Since most figures define poverty in terms of household income (e.g. household income per capita), reflecting the fact that families largely consume together, the information base is far weaker

A conceptual framework

on intra-household inequality, and this leaves a major challenge of thinking about and measuring the differences in poverty between males and females (in the same families), between adults and children etc. These are matters greatly affected by cultural patterns, among other things. 5 Lustig (2010) discusses the challenges of the present world recession from the perspective of protecting poor people in Latin America. 6 A good review of how insurance issues can be dealt with is provided by Perry et al. (2007: 182–183). 7 ‘Social insurance’ is a term usually used to refer to joint provision by the individual and the state. 8 The term ‘redistributive social protection’ might be used for such policy. 9 A good example of this was the evolution of the BRI banking system in Indonesia, leading to an enormous increase in financial savings by families of low or modest means (Papanek, 2004). 10 Lists vary with the author. The Asian Development Bank distinguishes five major kinds of activities: labour market policies and programmes, social insurance programmes, social assistance, micro and area-based schemes, and child protection. (A fuller breakdown of these categories is shown in Appendix B, Table 1.) 11 This result is assured if the equilibrium market wage would be below the mandated minimum wage. 12 Instead of making the contributions at each point in time during the worker’s career as an employee, the firm could simply have paid the money to the worker. In cases where the worker would then have consumed those additional wages, then the impact of the firm’s withholding them and in effect distributing them later is that the worker is de facto saving, with those ‘forced savings’ being managed by his/her employer. 13 Externalities are positive or negative effects from the behaviour of one economic agent (e.g. a consumer or a firm) on other agents, which are not accompanied by any payment in the other direction. Thus pollution by a firm is a negative externality if the firm does not have to recompense the victims. 14 A Conditional Cash Transfer programme in Mexico started in the late 1990s. 15 Such fees had been introduced by many countries in the 1980s and 1990s in an effort to bring in more funds to accompany the sometimes dwindling public resources available for health. But dramatic declines in service use ensued, particularly among vulnerable groups, while the frequency of catastrophic expenditure increased (WHO, 2008: 26). 16 However, their receipt of welfare may discourage work by family members, so in fact the end of the spectrum is the situation in which neither the beneficiary nor his/her family members are able to work. 17 Such scarring effects have been statistically identified mainly in industrial countries, but there seems no reason to doubt their presence in developing countries as well. 18 Maloney (2004), among others, states that the movement in both directions between formal and informal employment is much greater than some authors have believed. 19 The traditional systems were designed under the implicit assumption that workers would stay in the formal sector for their whole careers. In fact there is much movement in and out; in Mexico such movement has been shown to be greatest among low- and high-wage workers (Perry et al., 2007: 195). At present workers must accrue 25 years in the Mexican system to receive a minimum pension guarantee, but, given their non-continuous employment in the formal sector, it would take the average low-wage

Social protection, growth and employment worker 50 years to qualify, making it unattainable. Uruguay has a similar pattern, which is again a much bigger problem for low-income workers. Buchelli et al. (2006) estimate that about 50 percent of 60-year-olds with highest-quintile earnings will achieve the 35 years of service needed to receive a minimum pension, but almost no one from the lowest quintile will.

20 This is a major reason why some people do not contribute to health insurance in Colombia. Many females joining the labour force opt to stay out in Argentina for the same reason (Perry et al., 2007: 195). 21 Although by no means a case of no net benefits, Peruâ&#x20AC;&#x2122;s Integrated Heath System (SIS) shows the effects of inadequate pre-planning for the rapid growth of effective demand for its services. For a general discussion of such problems, see Grosh et al (2008).

A conceptual framework

References Auerbach, P., M. E. Genoni and C. Pagés. “Social Security Coverage and the Labor Market in Developing Countries,” IZA Discussion Papers 2979, Institute for the Study of Labor (IZA), 2007. Baeza, C. and T. Packard, Beyond Survival: Protecting Households from Health Shocks in Latin America, World Bank, Washington, D.C., 2006. Berry, A. “Employment and Efficiency Effects of Social Security (SS) and Social Protection (SP) Systems in the Context of an Informal Sector and Market Imperfections: A Conceptual Review”, mimeo 2013. Berry, A. and D. Mazumdar “Small Scale Industry in East and Southeast Asia: A Review of the Literature and Issues”, Asian Pacific Economic Literature, 1991 pp. 35-67. 1991 Berry, A. and J. Serieux (2007) ““Riding the Elephants: The Evolution of World Economic Growth and Income Distribution at the End of the 20th Century” in K.S. Jomo (with Jacques Baudot), Flat World, Big Gaps. London and New York: Zed Books, Orient Longman and TWN, in association with the United Nations, pp. 74-98, 2007. Berry, A., To Formalize or Not to Formalize: Trying to Ask the Right Questions, Manuscript, 2010. Birn, A., Y. Pillay and T. Holtz, Textbook of International Health: Global Health in a Dynamic World, Third Edition, Oxford University Press, New York, 2009. Buchelli, M., N. Ferreira-Coimbra, A. Corteza and I. Rossi, El Acceso a la jubilación on pensión en Uruguay; cuantos y quienes lo lograrían?, Economic Commission for Latin America and the Caribbean, Montevideo, 2006. Burgess, R. and R. Pande, ‘Do Rural Banks Matter? Evidence from the Indian Social Banking Experiment’, American Economic Review, 2005, 95(3): 780–795. Dollar D. and A. Kraay”Growth is Good for the Poor” Journal of Economic Growth, Vol. 7, pp.1195-225. 2002. Edmund Amann, Nektarios Aslanidis, Frederick Nixson and Bernard Walters. “Economic growth and poverty alleviation: a reconsideration of Dollar and Kraay.” European Journal of Development Research 18, no. 1: 22-44. 2006 Gasparini, L., F. Haimovich and S. Olivieri, Labor Informality Effects of the Programa Jefes de Hogar, Centre for Distribution, Labor and Social Studies (CEDLAS), Buenos Aires, 2006.

Social protection, growth and employment

Gavan, J.D. and I. Sri Chadrasekera, The Impact of Public Foodgrain Distribution on Food Consumption and Welfare in Sri Lanka, International Food Policy Research Institute (IFPRI), Washington, D.C., 1979. González de la Rocha, Mercedes., M.artha Moreno Pérez and I.nés Escobar González, ‘(2012) “Trabajo, Modos de Subsistencia y Vida Social en México’”, Conferencia Intérnamericana de Seguridad Social: Documento de Trabajo, CISS/WP/1205., Conferencia Interamericana de Seguridad Social/Inter-American Conference on Social Security, Mexico City, 2012. Grosh, M., C. del Ninno, E. Tesliuc and A. Ouerghi. For Protection and Promotion: The Design and Implementation of Effective Safety Nets. Washington: The World Bank. 2008. Hujo, K. and S. McClanahan (editors) Financing Social Policy: Mobilizing Resources for Social Development. Basingstoke: UNRISD and Palgrave Macmillan. 2009 Jalan, J. and M. Ravallion, ‘Are the Poor Less Well-Insured? Evidence on Vulnerability to Income Risk in Rural China’, Journal of Development Economics, 1999, 58: 61–81. Kakwani, N., S. Khamdker and H. H. Son, “Pro-poor growth: concepts and measurements with country case studies”, IPC Working Paper 1, August 2004. Khan, A., ‘“Employment policies for poverty reduction’,.” Issues in Employment and Poverty, Discussion Paper, No. 1, ILO, Geneva, 2001. Levy, S., Good Intentions, Bad outcomes: Social Policy, Informality, and Economic Growth in Mexico, Brookings Institution Press, Washington, D.C., 2008. Lindert, K., E. Skoufias and J. Shapiro, ‘Redistributing Income to the Poor and the Rich; Public Transfers in Latin America and the Caribbean’, Social Protection Discussion Paper, No. 0605, World Bank, Washington, D.C., 2006. Lustig. N., ‘Protecting Latin America’s Poor During Economic Crises’, Policy Brief, No. 2, Inter-American Dialogue, Washington, D.C., 2010. Maloney, W., ‘Informality Revisited’, World Development, 2004, 32(7): 1159–78. Miller, S., R. Lal, M. Lieuw-Kie-Song and D. Kostzer, ‘Employment Guarantee Programmes: A Response to Economic Crises and Structural Unemployment in Developing Countries’, Poverty Reduction Discussion Paper, No. PG/2009/003, UNDP, New York, 2009. Naudé, W. and A. Santos-Paulino (eds), Vulnerability in Developing Countries, United Nations University Press, Helsinki, 2010. Naudé, W., A. Santos-Paulino and M. McGillivray, ‘Measuring Vulnerability: An Overview and Introduction’, Oxford Development Studies, 2009, 37(3): 183–191.

A conceptual framework

Osmani, S. R., The Employment Nexus between Growth and Poverty: An Asian Perspective, A Report prepared for the Swedish International Development Agency (SIDA), Stockholm, and the United Nations Development Programme (UNDP), New York, 2004. Pagés, C. ‘Informality, Welfare and Productivity’, presentation at the Tercera Conferencia de Economía Laboral, Pontificia Universidad Católica del Perú, Lima, 29–30 Oct. 2009. Pagés, C., “A Cost-Benefit Approach to Labor-Market Reform.” Economic Review-Federal Reserve Bank of Atlanta, second quarter. No. 89, No. 2. 2004. Papanek, G.F., The Poor of Indonesia: The Impact of Economic Decline, Rapid Growth, and Crisis, 1952–2002, paper prepared as part of the Pro-Poor Economic Growth Research Studies, implemented by Development Alternatives, Inc. and the Boston Institute for Developing Economies, USAID, Washington, D.C., 2004. Patten, R.H. and J.K. Rosengard, Progress with Profits: The Development of Rural Banking in Indonesia, ICS Press, International Center for Economic Growth and the Harvard Institute for International Development, San Francisco, CA, 1991. Perry, G. E., W. F. Maloney, O. Arias, P. Fajnzylber, A. Mason and J. Saavedra-Chanduvi. Informality, Exit and Exclusion. World Bank, Washington, D.C. 2007. Rawlings, L. and G. Rubio, ‘Evaluating the Impact of Conditional Cash Transfer Programs’, World Bank Research Observer, 2005, 1: 29–55. United Nations Research Institute for Social Development (UNRISD). Combating Poverty and Inequality: Structural Change, Social Policy and Politics. UNRISD, Geneva. 2010. United Nations, Rethinking Poverty: Report on World Social Situation 2010 Media Kit, United Nations, Department of Economic and Social Affairs, Division of Social Policy for Development, New York, 2010. World Bank, World Development Report 2013: Jobs. Washington, D. C. 2012. World Health Organization (WHO), The World Health Report: Primary Health Care — Now More Than Ever, World Health Organization, Geneva, 2008. Zepeda, E., ‘“Pro-poor growth: What is it”’, IPC-IG One pager, No. 1, International Poverty Policy Centre for Inclusive Growth, Brasilia, 2004.

2 Social Protection and Labour Markets in Malawi: the Centrality of Agriculture Andrew Charman Introduction Malawi is a least developed country (LDC) with a predominantly rural population engaged in agriculture. The agricultural economy is dominated by small-scale farms (tended by people known as â&#x20AC;&#x2DC;smallholdersâ&#x20AC;&#x2122;) that produce commodity crops (mainly tobacco, coffee and cotton) and food for subsistence (maize, pulses, roots and tubers). Most smallholders cultivate less than one hectare. Poverty is deep and widespread across the land. As a result of their weak asset base, low technology adoption, limited land and labour constraints, the majority of rural households are highly vulnerable to shocks, whether generalized (such as drought) or household-specific (such as death). This weak resilience has hindered the ability of smallholders to move out of poverty and of the country to develop rapidly. Despite these structural impediments, there is now quantifiable evidence that measures intended to reduce smallholder vulnerability have begun to pay off, resulting in a decline in relative and absolute poverty. The benefit seemingly extends to all strata of poor people in Malawi, in urban and rural areas. Strengthening small farms has contributed towards rapid economic growth within the agricultural sector and broader economy, while at the household level it has led to improved food security as well as asset accumulation (cash, livestock and durable goods). This paper describes and analyses the impact that pro-poor programmes and social protection measures in particular have had on reducing poverty and mitigating the vulnerability of

Social protection, growth and employment

poor people in Malawi. Social protection here refers to the broad range of measures designed and implemented by the state and its partners in development to reduce poverty and strengthen the resilience of the population to shocks. The case focuses on the agricultural sector and its labour market dynamics. Over the past 15 years, a range of social protection measures has been instituted; the measures that have been most extensive in scale and scope, while having the greatest impact on poverty, are those that have sought to enhance the productivity of small-scale agriculture through input subsidies. The paper also examines the impact of these programmes and considers their potential to reduce poverty through providing decent work within the agricultural sector and through the role of agriculture as an engine of economic growth. The research for this paper was desk-based, drawing upon the latest secondary literature (including reports and surveys) and primary data on crop production from the Ministry of Agriculture and Food Security (MoAFS). The findings are informed by the important national surveys undertaken by the National Statistical Office (NSO), in particular the 2008 national population census (NSO, 2008a), the 2008 Welfare Monitoring Survey (WMS) (NSO, 2008b), the Finscope report (2009) and the earlier Integrated Household Survey 2 (IHS2) (NSO, 2005). The research uses the comprehensive Malawi Poverty and Vulnerability Analysis (MPVA) (GoM, 2006), which describes in detail the causes and characteristics of poverty. The paper also refers extensively to the findings of an International Labour Organization (ILO) funded study (Durevall and Mussa, 2010) that examined the countryâ&#x20AC;&#x2122;s labour market dynamics and recent economic performance.

Country profile Malawi is a central African nation of approximately 13.8 million persons. The country itself encompasses the lands of the southern and western flanks of Lake Malawi and much of this territorial water. Malawi is surrounded by the United Republic of Tanzania to the north, Mozambique to the east and south and Zambia to the west. Its topography has been fashioned by the forces that created the Central African Rift Valley. The main features include deep, wide valleys in the south that follow the course of the Shire River (which drains the Lake via the Zambezi River) and also characterize the Lake shore itself; highland plains that extend from the south to the north; and a high-altitude escarpment, whose features are most prominent in the north. The landscape reveals considerable diversity and is characterized by varying microclimates and ecological typologies. Across the territory, a range of farming systems has evolved to suit the varying agroecological conditions. The literature on vulnerability in Malawi places considerable emphasis on the influence of climatic factors, especially drought, as a cause of crop failure leading to food insecurity. However, the idea that drought is recurrent and severe is overstated. Rainfall data from the Directorate of Meteorology shows that, over the extended period of 1984â&#x20AC;&#x201C;2007, the national average was 1058 millilitres (1028 millilitres in the Northern Region, 1110 millilitres in the Central Region and 1039 millilitres in the Southern Region), whereas the lowest recorded average rainfall was 737 millilitres in 1991â&#x20AC;&#x201C;1992 (available from the Reserve Bank of Malawi (RBM) website). Malawi is indeed well watered in most districts and the respective

Malawi

agro-ecological conditions favour a diversity of commodity and grain crops. However, these various environmental conditions are not equally suited to the cultivation of tobacco and maize, the two crops at the forefront of the national political and development discourse. The population is predominantly rural. The urban areas accommodate the vast majority of the formal labour force (which is dominated by civil servants) and the fast-growing informal labour force engaged in microenterprises and unregulated trading (Durevall and Mussa, 2010). Malawi’s political and economic elite, along with the middle class, is urban-based. The most recent census, undertaken in 2007/2008, found that 85 percent of the population live rural lives and are engaged in agriculture and other livelihoods activities that use natural resources. The urban population is distributed between four cities and 27 district centres. The latter can best be described as no more than small towns that provide an administrative function and service industries. The geopolitical structure of the country strongly reflects its colonial heritage. Malawi became independent from the United Kingdom in 1964. The country comprises three administrative regions (the South, the Centre and the North) whose boundaries were inherited from the colonial era. There are six major tribal groups (of 500,000 persons or more) and another half dozen minor ones (NSO, 2008a). The independent government sought to shift the administrative heart of the country from the Southern to the Central region through positioning the capital at Lilongwe to promote greater regional integrity and mitigate political organization along lines of tribe. The power of tribes and their subsidiary clans remains strong, however, most notably in land administration (as most land is under customary tenure) and in supporting the administration to implement development programmes and social protection measures. In developmental terms, Malawi is classified as an LDC and is classed in terms of the UNDP human development index (HDI) as having a low human development—in other words, as having an HDI score below 0.5.1 In 2007, the most recent period for which data is available, Malawi had an HDI score of 0.493, which reflects the relatively low level of life expectancy at birth of 57 years, an adult literacy rate of 71.8 percent and a gross domestic product (GDP) per capita of US$761. On the HDI 2007 ranking, Malawi was positioned at 161 of 182 countries for which comparative data was available. Against UNDP’s Human Poverty Index (HPI-1), which takes into consideration the number of people living below the poverty measure of US$1.25 per day, Malawi had an HPI-1 value of 28.2 percent (or roughly one third of the population subsisting on less than this amount daily), placing it in rank 90 among 135 countries.

Demographic profile The 2008 census recorded a population of 13.07 million persons, an increase of over nine million persons since independence (NSO, 2008a). The population is currently increasing very fast; the inter-censual annual growth rate for 1998–2008 was 2.8 percent, whereas that of the previous period was 2 percent. Population density per square kilometre is 139 persons, though distribution is uneven and the density exceeds 200 persons per square kilometre in some districts, including

Social protection, growth and employment

those adjoining the major cities, whose populations have expanded at a very fast rate in excess of 4 percent per annum. Over time, the regional population distribution has shifted from the South towards the Central Region, which has now become the most populous and where the capital city, Lilongwe, is the epicentre of this rapid expansion. The rate of urbanization has surprisingly slowed since 1998, though. This trend reflects the comparatively better economic performance of the agricultural sector, and of small farms in particular, versus the manufacturing sector during this timeframe.2 The age structure of the Malawian population is highly skewed, with 48 percent under 15. The median population age is 17. The dependency ratio (0- to 14-year-olds plus over-65-yearolds to working population) has deteriorated over the last decade. The demographic structure of the population (urban versus rural), categorized in lustrums, is shown in Figure 2.1. The ratio of youth (under 18) to working adult population (18 to 64) will impact labour market dynamics once these persons reach working age. It is questionable whether the rural labour market will have the capacity to absorb all of these new workers in agricultural endeavours, let alone provide them with decent work; while the possibilities of being absorbed in other sectors are very small,

Figure 2.1: Population Age Structure, 2008

Source: NSO (2008a).

Malawi

given the concentration of formal sector employment in services, communications and mining, each of which demands highly skilled labour. Malawi has approximately 2.8 million households, comprised of immediate family members and relatives and sociologically ordered according to housing (NSO, 2008a). A household typically comprises several dwellings clustered together as an autonomous unit. The structure and composition of households frequently change, temporarily and permanently, through death, migration and resettlement and in response to exogenous and idiosyncratic shocks. The flexible composition of households reflects the strong ties of extended family and the culture of reciprocity, which, while strongest among families, extends to fellow members of the clan and in matters of politics across the entire tribe. The majority of households are male-headed, in cultural, though not necessarily legal, terms, as there are matrilineal and patrilineal customs amongst the different tribes. There are about 790,000 female-headed households, defined in sociological terms as the person whom members regard as the ‘head’, equivalent to about one third of all households. Female-headed households have come about as a result of separation, male migration and death. The census recorded more than 830,000 orphans (under 18 years). At the same time, only 67 percent of children live in a household with both parents; a number of children (11.73 percent) (roughly equivalent to the number of orphans) reside with other relatives or a single parent (predominately the mother). The population, as a whole, has received little education. Among persons older than five years, 2.6 million people (25 percent) have attended primary school, but only 249,019 persons (2 percent) have attended secondary school and a mere 8,877 people (0.08 percent) have attended university. At the time of the 2008 census, about 3.8 million persons in the over-five population were deemed illiterate, while only 18 percent of this same cohort was recorded as literate in the official (English) and the main non-official (ChiChewa) languages. Young male and female Malawians, however, are better educated than their parents, with the rate of literacy among 15- to 24-year-olds at 85 percent for males and 80 percent for females, though the goal of universal literacy remains elusive, as only 57 percent of children aged 10 to 14 are considered literate (NSO, 2008a). Apart from young persons attending school, the labour market participation rate is high. The 2008 WMS, for instance, reported that, among the adult working population (above 15 and below 65 years), labour force participation was 86 percent (96 percent among urban residents), implying that most Malawians were either working or actively seeking employment.3 About 3 percent of the population is recorded as being incapable of performing work due to infirmity. Most of those recorded as not working are in fact engaged in various household duties, including childcare. Among the economically active population, 88 percent of women and 72 percent of men reported working in agriculture for themselves (in other words, as self-employed smallholders). The WMS data (NSO, 2008b) reveals that 2 percent of men and 1 percent of women report their main source of remuneration as payment from casual labour or ‘ganyu’, as it is known locally. This is significant, as it confirms that the overwhelming majority of smallholders derive their main income from on-farm (own) agricultural activities and not paid labour, whether formal or informal. There are few formal employment opportunities within Malawi (with only 8 percent of the labour force receiving monthly payments in wages or salaries) and few nonformal opportunities outside agriculture.

Social protection, growth and employment

While the rural population is substantially engaged in farming activities, the demand for labour varies at different points in the season, resulting in labour shortages and seasonal or temporary unemployment. The MPVA Report (GoM, 2006: 41) argues that the main challenge of the labour markets is (seasonal) underemployment corresponding to the fluctuation in labour intensity across the agricultural season, which peaks during the periods of land preparation, planting, harvesting and post-harvest crop management. Scholars argue that, as a result of the high demand for labour during this time, poor people are often enticed (or compelled, if foodinsecure) to take up casual labour employment. Though this decision may have impact on their home production, the decision has been shown to be perfectly rational, as the returns from casual employment can exceed the on-farm productivity gain from tasks such as secondary weeding (Orr et al., 2009). The MPVA (GoM, 2006) estimates that approximately 5.2 percent of the adult male population can be considered to be time-poor, i.e. exhaustively engaged in agricultural endeavours (working more than 70 hours per week), whereas 15 to 20 percent of the same population are regarded as underemployed (i.e. working less than 10 hours per week on an annual basis). On average, working adults aged 15 years and above work varyingly between 33 and 43 hours per week across the year (ibid.: 42). Seasonal unemployment is considered to disproportionately affect individuals within the poorest quintile of the distribution of consumption income per capita (ibid.: 41). The MPVA report considers that, for adult women, seasonal underemployed is less evident due to the gendered nature of household labour, wherein women undertake the largest share of domestic work, including child rearing. This characterization of seasonal unemployment, though difficult to quantify, potentially undervalues the engagement of the rural population in non-agricultural and livelihood activities, such as house construction, natural resource harvesting, trade and family-related reciprocal duties that can sustain high rural labour demand and minimize unemployment.

Macroeconomic trends From the late 1970s to the mid-2000s, the Malawian economy grew slowly, with GDP per capita (in constant terms) fluctuating annually and the major downward swings of 1992, 1994 and 2001 corresponding to years of drought. Since 2004, the Malawian economy has grown rapidly. Annual production increases in maize, the countryâ&#x20AC;&#x2122;s main food crop, have contributed significantly towards GDP growth (Durevall and Mussa, 2010: 55) and, at the household level, had a positive income effect (see Dorward, Chirwa et al., 2008). Over the period 2004 to 2009, Malawi outperformed its LDC neighbours, Mozambique and Zambia, in maize productivity enhancement, performing nearly as well as the large-scale commercial farmers in South Africa. Smallholder in Malawi enhanced their production in response to market incentives (and took advantage of the favourable growing seasons). Data from FOASTATS shows that, over this time, Zambian maize production rose by 8 percent, Mozambican by 14 percent, Malawian by 58 percent and South African by 64 percent. Because the increases in Malawi correlate with a period of sustained input subsidization, this achievement correlates with a period of input subsidization, as did the similarly enhanced output achieved in the period 1998/1999 to 1999/2000. In

Malawi

both periods, the common denominator was the transfer of cheap fertilizer and higher-yielding seed to small farmers, though rainfall had an important role as well. GDP rose (in constant 2005 prices) from US$2.76 billion in 2005 to US$3.74 billion in 2009, a growth of 35 percent (RBM, 2009; 2010). Crop and animal production provide the largest subsector contribution towards GDP. Smallholders account for the bulk of agricultural output in all sectors apart from sugar, tea and flue-cured (Virginia) tobacco, although smallholders produce over 90 percent of burley (air-cured) tobacco. Smallholder output has significantly increased in this time and, although this growth has been uneven between seasons, the evidence over the long run shows a rate of expansion that exceeds population growth. The contribution of crop and animal production to GDP has risen from 27 percent in 2005 to around 32 percent in 2009 in constant terms. The second most important subsector is wholesale and retail trade, which accounts for roughly half of this level. The growth in the economy has translated into higher gross national income (GNI), which rose from US$170 per person in 2004 to US$280 per person in 2008. Over this period, inflation fluctuated on a downward trend to about 7.5 percent in 2010. Several economic drivers have contributed towards this seemingly impressive macroeconomic performance. Foreign direct investment rose substantially from 2004 onwards, led by a single mining investment (Malawi is poorly endowed in mineral resources). There were also significant improvements in the price of agricultural commodities, including tobacco (Malawi’s main export) and tea (Malawi’s third major export), thus bolstering the government’s revenue position. The production (in volume terms) of tobacco and tea reached a high point in 2009. Although some commodity prices (notably tea) are currently in a downward cycle, burley tobacco prices improved from 2007, benefiting government revenue and smallholder producers. Tobacco exports accounted for 67 percent and 65 percent of the total value of domestic exports in 2008 and 2009, respectively. In both years, agricultural commodities (counting tobacco, sugar, tea, coffee, cotton and legumes) accounted for approximately 80 percent of the value of exports. Malawi was afforded debt relief under the Highly Indebted Poor Countries/Multilateral Debt Relief Initiative in 2006 (International Monetary Fund (IMF), 2006). Debt relief has significantly relieved pressure on the revenue account and enabled the ruling wa Mutharika government to undertake far-sighted policy interventions, notably provision of productivity enhancement support to smallholders. Still, the country remains heavily dependent on direct donor funding and programmes. In this respect, the government’s revenue position has benefited from a continued inflow of grant funding into the health and education sectors, for food security/social protection measures, and as balance of payment support. Grant funds accounted for approximately 26 percent of total government revenue in 2005, 28 percent in 2006, 41 percent in 2007, 21.2 percent in 2008 and 17.2 percent in 2009 (RBM, 2010); the diminished figure in 2009 reflects the withholding of disbursements as donors awaited IMF approval of a new programme for Malawi. Foreign aid has increased from roughly US$500 million per annum in 2004 to US$900 million in 2008 (Durevall and Mussa, 2010: 90). The present government has largely abided by its commitments to the IMF to maintain fiscal discipline, adhere to budgets and maintain parliamentary oversight. Its record of prosecuting corruption within the civil service,

Social protection, growth and employment

preventing the subversion of development programmes by the political elite,4 and addressing the process of ‘quiet corruption’ that hinders doing business in the country is much less impressive.5 Grassroots opposition to the wa Mutharika government intensified from circa 2010 in response to foreign exchange shortages, rising fuel costs and high unemployment. In 2011 wa Mutharika expelled the British High Commissioner, as evidence of his criticism of the Malawi government became public via wikileaks, resulting in the suspension of donor aid. As the country’s economic situation rapidly deteriored from this point, the wa Mutharika government was confronted with nationwide protests and opposition from within the ruling party. Wa Mutharika passed away in April 2012 and was succeeded by the suspended vice president Joyce Banda. The impact of this crisis of leadership and allied fiscal instability on the agricultural sector and smallholder output is outside of the scope of this paper. The Malawian macroeconomic profile also contains structural weaknesses and expenditure dependencies. The economy relies heavily on agriculture. Shocks to production and profitability—such as drought, rising input prices or falling commodity prices—have a major impact on growth and livelihoods. Over the past two decades, the negative impact of drought on GDP is evident, especially in 1992, 1994, 2001 and 2004. The country has a large trade deficit, driven through growth in consumption and an increasing dependence on imported foods in certain categories (such as wheat) that the smallholder sector is unable to produce. Since 1994, local manufacturing capacity has weakened in all sectors, apart from the provision of electricity and water and mining. In the food, beverages and tobacco subsector, the index of industrial production (1984 = 100) plummeted from a level of 165.4 in 1994 to 68.4 in 2006. In contrast to manufacturing, the economy has been aided from rapid growth in mining and quarrying, information and phone communication, financial services and services relating to the provision of education and health care (RBM, 2010). While broadly committed to strengthening the market, the current government has sought to intervene in its operation at various levels to reduce price volatility and to improve market inefficiencies. In recent years, for example, it has adjusted upwards and sought to fix prices for strategic crops, including tobacco, maize and cotton. While these actions have in some instances achieved popular results, they have also had unintended impacts, such as causing buyers in the private sector to withdraw from markets (cotton) and raising transaction costs (on fertilizer distribution) (see Dorward, Chirwa et al., 2010). Private-sector agribusiness in most crop sectors, bar tobacco, is still weakly positioned, having been excluded from markets during the Banda era, when parastatals monopolized the supply of inputs and crop sales. The development of private fertilizer distribution into remote districts was gradual and free input and input subsidy programmes were seen (from industry perspective) as an opportunity to develop and a threat through which distribution was again monopolized by state companies. Another area of concern relates to the overvaluation of the kwacha against the US dollar, causing foreign exchange shortages (Durevall and Mussa, 2010: 84–86). Since 2006, the government’s monetary approach has pegged the value of the kwacha to the dollar through controlling foreign exchange sales to the private sector; the rationale has been to reduce inflation and prevent the occurrence of price spikes, which would happen if the kwatcha were to devaluate. The strategy has helped to reduce

Malawi

the cost of inputs (especially fertilizers) and enabled consumers to sustain their dependence on cheap imports of food and consumer goods. But the risks of these measures on long-term economic development are considerable. A recent ILO study warns that the overvaluation will ultimately harm domestic industry as a consequence of the natural resource curse, leading to job shedding, while the comparatively high rate of inflation will erode the profit on exports over time (on the pitfalls of this strategy, see Berry, 2008). The small size of the manufacturing sector, at this point, lessons the scale of the immediate danger from currency overvaluation.

Poverty As Malawi is a predominately rural nation of subsistence farmers, the character of poverty there has been shaped by the structural features of the rural economy: over 90 percent of households produce their own food on small landholdings using traditional farming methods; the majority of rural households derive their main source of income from crop sales; the population is very young and fast-growing, burdening households with a high dependency ratio; and poverty is deep and widespread, with the income gap between poor and non-poor people fluctuating seasonally so that households periodically move into and out of poverty. The World Bank and the Government of Malawi (GoM) commissioned in 2006 a stateof-knowledge investigation into poverty and vulnerability, drawing on the most up-to-date data from the NSO IHS2. One of its main conclusions was that â&#x20AC;&#x2DC;poverty is pervasive and not merely the situation of the lowest economic groupsâ&#x20AC;&#x2122;. It found that the poverty level (using NSO benchmarks) had not improved in the period between 1998 and the assessment (c. 2004). However, it recognized that poverty is a fluctuating dynamic, with two thirds of all households having at some point moved either into or out of poverty over the course of the decade after democracy. Their definition of poverty focused on total annual consumption expenditure per capita, taking into consideration an individualâ&#x20AC;&#x2122;s food requirements and other non-food consumption needs. It should be noted that the IHS definition is not weighted by age structure, which is significant, given the large share of the population that comprises children. Employing a regression analysis of per capita consumption, the MPVA identified these main correlates of poverty: household size, education, access to non-farm employment, access to irrigation, proximity to markets and trading centres and tarmac roads, landholding size and engagement in cash crop production. Female-headed households were correlated with lower consumption levels. It found, contrary to past assumptions, that the presence of orphans within the household had no pronounced influence on poverty levels, a finding that has since been confirmed in other studies (Devereux, Baulch et al., 2006: 53). It found that larger households are poorer than smaller households, with consumption declining by about 5 percent for each additional child. Households whose head had completed primary education had higher per capita consumption (12 percent on average), as had households that operated off-farm businesses. The average landholding size per household, as reported in the IHS2 survey, was 1.2 hectares, equivalent to 0.33 hectares per capita. Land is held under communal tenure and all land that is not cultivated is (in effect) available for common grazing; the main large livestock that use

Social protection, growth and employment

the commons are cattle and goats. There are significant regional variations in landholding size. Landholding size diminishes in the Southern Region, while individual landholdings are slightly larger in the Northern Region. The non-poor have larger landholdings; however, as the MVPA points out (GoM, 2006: xviii), even cohorts in the highest consumption deciles have relatively small landholdings, which are significantly smaller than the average size of tobacco estates (12 hectares to 20 hectares) held by the elites under leasehold title. While larger landholdings could be correlated with a slight improvement in household consumption per capita (4 percent per additional hectare) (one reason for the small increase is that larger landholdings include inferior arable land), access to wetlands/irrigated lands for winter cropping was considered to have an even more significant impact. These wetlands enable smallholders to produce vegetables and maize for household consumption without mechanical irrigation during the dry season (when rain-fed cultivation is not possible), which extends from May to October. The assessment found that the proximity of smallholders to trading centres and transport infrastructure is correlated with poverty, with household consumption declining over distance. These determinants have not changed significantly since 2004. What has changed, as we discuss below, is the scale of the influence of cash and food crops. The MPVA report notes that “cash crop production is an important path out of poverty”, while the per capita consumption of smallholders who produce tobacco is 6 percent higher on average. The MPVA report distinguished between ultra-poor, poor and non-poor households, a distinction made in the first IHS (NSO, 2005: 138). The socio-economic characteristics of these cohorts are described in Table 2.1. In the IHS2, poor households were defined as those whose annual per capita consumption expenditure was below US$0.50 per day (MK16,165 per person per year), which was considered to be an appropriate poverty line.6 Within the poor population, ultra-poor households were defined as households whose total annual consumption expenditure on food and non-food items was lower than the minimum food expenditure deemed necessary to sustain a person’s existence on the basis of the recommended daily calorie requirement. The survey set a poverty line of MK10,029 per person per year as the cut-off. The IHS2 determined that, in 2004, 52 percent of Malawian households were poor, while 22 percent of the total population of households—in effect, just under the half of poor people were ultra-poor. This state of household poverty at the national level mirrored the findings of the first IHS, undertaken in 1998, which estimated that 53.1 percent of households were poor and 23.6 percent were ultrapoor. In 2004, the average person had an income about 63 percent of the poverty threshold in terms of adequate per capita consumption requirements; the income shortfall, according to the MPVA, equated to about 16 percent of GDP (GoM, 2006: 234). The NSO reassessed the scale of poverty (in terms of consumption expenditure per capita) annually from 2005 to 2008 through its WMS. The 2008 WMS reports a reduction in the proportion of poor and ultra-poor people. It found that the number of poor households within the total population had declined on a yearly basis: according the latest survey, 40 percent of households were poor, with the rate of decline sharpest among the urban population. A reduction in poverty was also evident among households in the lowest income quintile (i.e. ultra-poor). It reported that the proportion of ultra-poor households had annually declined from 22 to 15 percent. This

Malawi

Table 2.1: Indicative Features of Poverty Ultra-poor

Poor

Non-poor

Scale c.1998â&#x20AC;&#x201C;2004

22%

30%

48%

Scale c. 2005â&#x20AC;&#x201C;present

15%

25%

60%

Household head

Female (as well as Male) Male (some Female)

Male

Land

No access to wetlands

Access to wetlands

Landless

Labour

Landed, but low productivity levels

Landed, but extremely low productivity levels

No or limited access to wetlands

Predominantly household

Household, plus extended family

Larger/more fertile landholdings Household, plus extended family Casual workers are employed in peak season

Assets

Minimal

Minimal 3% of households have oxcarts 62% of households have radios

Education

Household head limited schooling

Household head completed primary school

Household size

Large household

Smaller household

Geographical factors

Far from trading centres/ roads

Cropping focus

Food crops

Food crops, with some investment in cash crops

Maximum investment in cash crops, minor investment in food crops

Off-farm economic activities

Piece-work and resource harvesting

Piece-work and trading

Trading and small enterprises

Close to trading centres/ tarmac roads

Source: Based on MVPA (GoM, 2006) and IHS2 (NSO, 2005) findings

reduction in poverty was evident in all three regions, although the rate of decline in poverty has been comparably slowest in the Central Region, the heart of the tobacco economy. The reduction in poverty in the Southern Region can be least attributed to tobacco (only 9 percent of households grow tobacco), so the path out of poverty must owe to food crops (maize and pulses) and off-farm income. The absence of piece-work opportunities on tobacco farms may possibly account for the comparatively slower decline in ultra-poverty within this region. In contrast, the availability of these work opportunities in the Central and Northern Regions may explain the far lower proportion of households in the ultra-poor cohort who thus benefit from seasonal work.

Social protection, growth and employment

Table 2.2: Proportion of Households That Own Selected Assets, 2004 vs. 2008 Household Assets House (owned)

Bed

Table

Radio TV and Bicycle Ox cart (wireless) VCR

Hoe

Axe

IHS2 Urban

Rural

1st quintile

2nd quintile

3rd quintile

4th quintile

5th quintile

Urban

Rural

WMS, 2008

Source: IHS2 (NSO, 2005); WMS (NSO, 2008b)

The IHS and WMS collected data on assets that, although not comparable across expenditure quintiles, reveal a narrow gradient in asset ownership between poor (1st quintile) and non-poor people (5th quintile); the data are shown in Table 2.2. In most of the assets under consideration, the difference between poor and non-poor people is negligible. The WMS reveals that the household asset position of the rural population improved from 2004 to 2008 in all assets (apart from axes); non-productive assets (furniture, radio/TV and bicycles) increased the most, improving from a low base. The number of households with ox carts (a proxy for farm productivity) remains very small and highlights one of the obstacles to further enhancing agricultural output and accessing markets. The low level of farm technology presents a similar hurdle. Apart from assets, most smallholders have some form of savings, although most do not bank but retain cash reserves. The Finscope survey (2009) found that 74 percent of adult respondents were saving either in cash (62 percent) or in kind (38 percent), such as investing in livestock or purchasing farm inputs. The respondents said that their main reason for saving was to reduce the impact of shocks, especially illness and death. Despite their intention to save, 80 percent of the respondents claimed that there were times of the month during the year, especially towards the end of the dry season, when there was no cash in the household, a finding that highlights the high levels of vulnerability among poor people. The 2008 WMS reported an improvement in the nutritional status of children under the age of five since 2004. This achievement is partly attributable to increased household food availability and partly to the scaling up of nutritional programmes. The report shows that the proportions of underweight, stunted and wasted children have decreased. There remain major challenges in terms of food insecurity, poor nutrition and low income. The MPVA deduced that the cause of

Malawi

malnutrition could not be attributed to the depth of poverty alone, as chronic malnutrition was observable in all consumption quintiles. Moreover, landholding size was not demonstrably positively correlated with malnutrition, while less stunting was found in female-headed households. Poor nutrition reflects inadequate diet (and micronutrient deficiency), a problem that seems most pronounced in households of undiversified farming systems. Despite recent improvements in national maize availability, there are still food-insecure households, some as a result of livelihood shocks specific to certain geographical localities or the food price spikes that precede each maize harvest. But the weight of evidence suggests that household resilience across all wealth quintiles—excluding the destitute—has strengthened.

Vulnerability The periodic vulnerability of rural Malawians to hunger and chronic food insecurity has attracted global attention, most recently in 2001 and 2002, when hunger was widespread. There is an extensive body of literature on this topic. The prevailing view is that poor people in Malawi are caught in a worsening ‘poverty trap’ (Devereux, Baulch et al., 2006; World Bank, 2007).7 It is thought that the vulnerability of poor people has, over time, been ratcheted upwards (Dorward and Kydd, 2004); the ratchets include failure of markets, labour scarcity, land shortage and a rising household dependency ratio. One scholar contends, “Malawians are more vulnerable today than in the past because hazards appear to have increased—rainfall and food production are erratic, HIV/AIDS is spreading, markets are weak and prices are volatile—and their ability to cope has declined—livelihoods are dangerously undiversified, repeated shocks have eroded assets and savings, informal networks are less willing or able to provide assistance” (Devereux, Baulch et al., 2006: 14). Livelihood shocks occur annually to some portion of the population, as a result of their geographic location (situated in a flood or drought locality) or impoverished position and thus inability to overcome a minor deviation in, for example, rainfall distribution. These shocks are referred to as covariate shocks (following the tradition set by Binswanger and McIntire, 1987). Households are also vulnerable to individual or idiosyncratic shocks that affect the family alone, such as death, disease or indebtedness. Shocks to smallholder livelihoods, whether covariate or idiosyncratic, can cause non-poor households to move ‘in’ to poverty or poor people to become ultra-poor for as long as it takes them to recover their position. The process of moving in and out of poverty is now well recognized in the Malawian literature (see GoM, 2006; Devereux, Baulch et al., 2006). The IHS2 sought to quantify the extent of vulnerability within the population. The survey found that the majority of households had experienced some form of severe livelihood shock in the preceding five years. Table 2.3 details the most commonly reported shocks. There is a general consistency in the nature of the shocks experienced across income quintiles (GoM, 2006: 60). These shocks affect households in two respects. One set of shocks, those relating to prices and production, undermined the ability of the household to meet its food needs, which, in turn, affected its welfare and income. The other set of shocks, those relating to illness and

Social protection, growth and employment

Table 2.3: Proportion of Households Affected by Severe Shocks, c. 2004 Shock

All

Urban

Rural

Maleheaded

Femaleheaded

Rise in food price

Reduction in crop output (drought/ flooding)

Illness of household member

Death of family member

Fall in the price of crops

Crop disease/pest damage

Livestock loss

Source: IHS2 (NSO, 2005:136)

death, had a direct impact on household income, requiring an expenditure outlay. It should be recognized that the original survey was undertaken at a time of declining maize production, falling tobacco prices and fluctuating maize prices. Household vulnerability to food insecurity had intensified. At this time, household concerns over health were also at a high point because HIV prevalence seemed to be on an upward trajectory. The fear of death and illness at that time would be understandable. A more recent survey undertaken in 2007 found that most of these shocks remained relevant. But a new concern now topped the rankings: about one third of the respondents reported being affected by the shock or stress of low crop yields due to poor soil fertility (see Dorward, Guenther et al., 2008). This can be interpreted as their uncertainty of access to inorganic fertilizers. The IHS2 survey investigated the strategies that households had used to overcome shocks. Sixty percent of households did nothing at all, but managed to survive the crises without undermining their livelihoods. The most common coping strategy (37 percent) was to intensify their labour on-farm or elsewhere. About one quarter of households survived through reducing consumption and adopting asset-depletion strategies such as spending cash savings, selling assets and animals and borrowing money from relatives (NSO, 2005: 137). The MPVA reports that the most common ex ante strategies at this time were income diversification, crop diversification and migration (GoM, 2006: xxvi).8 Labour migration has historically been an important strategy for poor people to cope during crisis (livelihood shocks) and a medium-term strategy to accumulate household assets and income. The main regional destination for Malawian migrants is South Africa. Despite improvements in national food production since 2004, the threat of food insecurity remains a major source of vulnerability for the rural population, especially those households that are (partially) reliant on food purchases. Throughout the past decade, the price of the main food staple, maize, has exhibited a seasonal fluctuation, responding to supply and demand conditions (MoAFS market price data). In the months after August, when household food supplies begin to run low, to January, when the first green maize of the new season enters the market,

Malawi

food-insecure households are vulnerable to the shock of steep rises in maize prices. A sharp increase in maize prices will affect the ability of households, especially that of poor households, to meet the balance of their food requirements from conventional markets. If other sources of food are not obtainable through non-formal markets and intra-family reciprocity, this situation can result in a shortfall in household food entitlements, ending in hunger and nutritional deficiency. Maize prices in 2008, for example, more than doubled from US$0.23 per kg in May to US$0.50 in February 2009 before the first new crop entered the market. The Malawi Vulnerability Assessment Committee (MVAC) has identified tens of thousands of households as being food-insecure—in terms of missing household food entitlements—in the two previous seasons. In the 2008/2009 season, MVAC quantified the food-insecure population as up to 275,000 persons, whereas, for the 2009/2010 season, the MVAC determined the figure to be 1.1 million persons, or 8 percent of the population (FEWSNET, 2009; MVAC: Vol. 6, No. 1). In both seasons, their findings have led to a call for a resumption and intensification of food aid in the form of free food distribution and other measures to provide a more substantial ‘safety net’.

Social protection Over the past two decades, initially in response to the strategic recognition that structural adjustment policies could not redress poverty (see Harrigan, 2003), measures have been instituted to reduce the vulnerability of the rural population to external shocks. The evolution of the current array of social protection interventions was influenced from two directions: from one side, policies were shaped by the necessity to implement safety net measures, especially food aid, in support of humanitarian relief; from another side, the policies and programmes were shaped by the long-term development objective of enhancing growth within the agricultural sector and, in particular, of improving the capacity of smallholders. On both sides, protagonists contested the nature of social protection and its role in reducing poverty and strengthening resilience, resulting in frequent changes to policy and the rearticulation of programmes. The contestation over social protection has continued into the present time and it is therefore important to locate current programmes within their historical and political contexts.

Historical shifts in development policy Agricultural policy and government strategies for rural development have undergone significant changes since independence. The major institutional actions that have shaped social protection in Malawi over time are summarized in Table 2.4. While these shifts reflect various other economic and political influences, most policies have been influenced (to different degrees) by the president, first President Banda (1964–1993), who led the country to independence, then President Mluzi (1994–2004), who spearheaded the transition to multi-party democracy, and President wa Muthrika (2004–2012). The changes in policies and refinement in strategies are thus categorized by presidency. Throughout this period, the shifts have been affected by three ideological considerations. First, there was the gradual recognition (among the Malawian political class) that the position of independent small farmers is central to agricultural growth and

Social protection, growth and employment

poverty reduction, which required a redirection in policy emphasis away from large enterprises and resettlement schemes. Second, there was the gradual recognition (among the donor community) that neoliberal remedies could neither achieve agricultural growth nor poverty reduction and that direct support to small farmers was necessary to achieve green revolution goals. Third, there was the contestation between government and donors (and among donors themselves) over

Table 2.4: Timeline of Key Institutional Actions for Social Protection 1964

Malawian Independence

1968/1969

Integrated Rural Development Projects in Lilongwe, Shire Valley and Karonga districts

1970s–1990s

National Rural Development Programme; first of four World Bank-funded programmes

1971

The Agricultural Development and Marketing Corporation (ADMARC) established by an Act of Parliament and mandated to market inputs and crops for smallholders.

1986

State of Development Policies (DEVPOL); revised rural development strategy for the period 1987–1996, which ended in 1995

1988

The Smallholder Agricultural Credit Administration (SACA) was established as an organization within the Ministry of Agriculture to supply smallholders with agricultural credit; scheme collapses c. 1992

1994

Change of government: United Democratic Front

1994

Malawi Rural Finance Company established as an independent legal entity to provide farming input credit to smallholders, thus replacing SACA

1995

Agricultural and Livestock Development Strategy and Action Plan (ALDSAP) launched

1995–present

Malawi Social Action Fund (MASAF) established as the major social protection instrument to mitigate social aspects of adjustment, funded through World Bank loans

1998–2000

Government implements Starter Pack Programme

2001–2004

Government implements the Targeted Inputs Programme (TIP)

2000

Government launched the National Safety Nets Strategy

2002

Government initiates a National Safety Nets programme under the World Banksupported Malawi Poverty Reduction Strategy (MPRS)

2004

Change of president: Dr. Bingu wa Muthrika

2005

Government implements the Agricultural Inputs Subsidy Programme (AISP)

2006

The Malawi Growth and Development Strategy (MGDS) 2006–2011 launched as the successor to the MPRS. Policy specifies objectives to achieve social protection and disaster risk management as a specific developmental theme.

2006

Government commences the Agriculture Sector Wide Approach (ASWAp)

2009

Government launches a Social Support Strategy

Malawi

policy, with donors seeking to micro-manage programmes to achieve equity and efficiency, but causing ongoing programme redesign and reversals. The Banda era (1961–1994) Upon Malawi’s gaining of independence, President Banda sought to redress the marginalization of the indigenous peoples and serve the class interests of the ruling Malawi Congress Party’s (MCP) main constituents, comprising civil servants and urban wage earners; while urban, both stakeholders held strong interests in farming. There were four main aspects to the strategy he directed to achieve this objective.9 First, the government sought to develop large-scale agricultural schemes to enable smallholders to produce crops under intensive farming systems; these were supported by World Bank-funded Rural Development Projects. Second, it alienated large landholdings that were then awarded to parastatal agribusiness entities to produce select commodities (tea and tobacco). Third, it sought to provide opportunities for the middle class and political elites to acquire private landholdings (‘estates’ under 99-year lease hold tenure) on which to produce tobacco, a crop that smallholders were not permitted to grow. The area under estates grew from about 79,000 hectares in 1970 to about 1,180,000 hectares in 1998 (Chilmampunga et al., 1998: 4). Fourth, the state sought to restrict private-sector opportunities to trade consumables and agricultural commodities, allowing parastatals (notably ADMARC) to dominate input and output markets. The Banda strategy was successful in fostering rapid economic growth. In the period between independence and the late 1970s, the economy (GDP) grew at 5.5 percent per annum, propelled through the rapid expansion of the estate sector (which grew 17 percent per annum) and gradual reorientation within the smallholder sector towards commodity production (which grew at 3 percent per annum) (Harrigan, 2003: 848). By the late 1970s, the government had settled on an overarching strategy for promoting growth within the smallholder sector, as articulated in the first DEVPOL. The strategy was based on the transfer of green revolution technologies (higher-yielding seeds as well as mechanization), the establishment of a pan-territorial infrastructure to provide agricultural extension services, credit support for commodity producers, and market access through ADMARC. The success of this strategy was to be measured, politically, in terms of maize production. The World Bank channelled funding (approximately US$179 million) towards these objectives through successive National Rural Development Programmes. While the Breton Woods institutions had bought into aspects of this strategy, it simultaneously used Structural Adjustment Loans (SAL) to reduce the role of the state and promote price reforms, aggressively pursuing a neoliberal agenda. The country obtained SALs in 1981, 1984, 1986 and 1988 in a period that one scholar has described as enforced adherence to the principles of “pricism and state minimalism” (Harrigan, 2003). The most contentious aspect of these reforms was the requirement for the staggered removal of fertilizer subsidies, to which the government grudgingly agreed. At this time, fertilizer subsidies primarily benefited the estate sector and non-poor smallholders, largely those organized in commodity club structures (cotton, coffee). The second DEVPOL commenced in 1987. The new development strategy was set against the backdrop of a major production shortfall in maize and stagnating national economic growth,

Social protection, growth and employment

due in part to the impact of the civil war in Mozambique. Unlike in the earlier DEVPOL, the government in this phase was better able to resist the pressure for further liberalization, not least because strategists at the World Bank had begun to recognize that the poor sequencing of price reforms and lessoning of state support for smallholder agriculture had, in fact, hindered growth (see Harrigan, 2003: 849). In 1987, under President Bandaâ&#x20AC;&#x2122;s direction, the government reintroduced the provision of fertilizer subsidies (22 percent below market price) and increased maize prices (raised by 36 percent). It was subsequently agreed to prolong the process of subsidy removal. According to the terms of the fertilizer subsidy removal programme under the 1998 SAL, smallholder producers were expected to meet the full costs of inputs by 1998. In order to cushion the anticipated negative impact, the government sought to extend seasonal credit from 16 percent of smallholders to between 25 percent and 33 percent of all smallholders through the mechanism of the Smallholder Agricultural Credit Administration (SACA). SACA initially succeeded in this goal, though its fortunes were tied to those of the ruling MCP and the service collapsed in 1993 (on the eve of political transition) as a consequence of wide-scale loan defaulting. A more pro-poor agricultural development strategy began to emerge, ironically, as the power of the Banda era started to wane. The need for smallholders to be afforded more substantial support and better access to markets was underlined during the 1992 drought when maize production fell to a low point and hunger was widespread. The resulting humanitarian crisis was mitigated through the provision of food aid and distribution of agricultural inputs packages (maize seed and fertilizer) to smallholders. At the same time, the government was pressured by the IMF/World Bank to undertake further liberalization reforms, including (most significantly) reduced barriers for smallholders to grow tobacco. The Mluzi era (1994â&#x20AC;&#x201C;2004) The MCP was defeated at the 1994 multi-party elections. Bakili Mluzi, a wealthy businessman and trader, headed the new government. The political heartland of the United Democratic Movement (UDM) and its allies were poor people in urban areas and marginalized smallholders in the Southern and Northern Regions. It had no allegiance to the urban elites and the tobacco estate owners who had become the main power base of the MCP. The UDM sought to appeal to these constituents through a populist, pro-poor rhetoric and by instituting policies that would provide modest welfare gains for all Malawians in primary health and education as well as transfers of agricultural inputs and free food. The Mluzi government has been described as pursuing a strategy of neo-patromonalism through financially rewarding its clients and supporters. Its failings in governance are well noted (Chinsinga, 2007); during this era, corruption became pervasive within the civil service, while their performance in carrying out their duties deteriorated quickly. The governmentâ&#x20AC;&#x2122;s control of the macroeconomy was weak, resulting in rapid inflation and accumulating debt. Similarly, the senior leadership lacked capacity (and will) to shape policy, allowing donors to determine key policies for agriculture. This era has been described as a period of policy discord. Despite these criticisms, the UDM leaders were politically attuned to the welfare needs of poor people in rural areas and sensitive, in particular, to their ambitions to move out of poverty and live

Malawi

better than their forefathers had under President Banda. The mood for self-improvement was captivated in the government’s support for Vision 2020 (National Economic Council, 1998), a strategy adopted in 1999. Within the framework of this vision, ordinary Malawians desired to see themselves better clothed, better educated, better advanced and wholly non-poor. The Mluzi government’s initial strategic intention was set out in the ALDSAP. This development plan sought to overcome the dualism in Malawian agriculture through positioning smallholders at the centre stage of agriculture support services (MoALD, 1995). Its main pillars were to enhance maize productivity, significantly extend opportunities for smallholders in tobacco, and reform institutions, thus providing greater scope for private-sector involvement in input and output markets. As a consequence of these changes, smallholder output in tobacco and maize increased at a significant rate after 1994 (although the latter with unevenness), whereas estatesector output went into decline. While smallholders did benefit from market openings, the scale of government support towards the sector did not match the ALDSAP objectives. From 1986 to 2002, expenditure on agriculture from the revenue account (i.e. recurrent expenditure on operational costs) steadily fell from a high of US$430 million in 1989/1990 to US$373 million in 1998/1999, while expenditure from the development account (i.e. investment in programmes and infrastructure development) fell significantly between 1994 and 1998 (Charman, 2004: 46–53). The decline in investment was felt most sharply at the Agricultural Development Division (ADD) (i.e. subregional) level, where specialist services were effectively discontinued. Moreover, investment in the main institution responsible for training field extension officers decreased from US$962,254 in 1990/1991 to less than US$250,000 in the late 1990s, with the college then closed down temporarily. Against this backdrop of sharply declining resources, the donor community was able to strengthen its hand in policy-making, using the resources they could muster as bargaining chips. Weak leadership within the Ministry of Agriculture permitted numerous donor programmes (notably those of the European Union (EU), the United Kingdom (DFID), the United States of America (USAID) and Japan) to pull in different directions, each seeking to shape and influence agricultural policy in its own fashion. Disagreement over policy between government and donors (and among the donors themselves) continued throughout the Mluzi era. One of the main areas of disagreement concerned the role that parastatal bodies should fulfil in the agricultural sector, including the function of ADMARC, the National Food Reserve Agency (NFRA) and the Smallholder Farmers Fertilizer Revolving Fund Mechanism (SFFRFM), all of which government sought to sustain, but, lacking the financial means, was unable to do. The donors agreed that these institutions (notably the food reserve) were costly to maintain and that the emerging private sector could fulfil its role in input/ output markets more effectively and with greater cost benefits for consumers. The discord was tempered after 2001 as a result of a moderate food crisis, when maize production fell and the institutional mechanisms intended to ensure maize availability and household accessibility were found wanting. The experience of the subsequent 2001–2003 period demonstrated the case for maintaining a strategic grain reserve and the ADMARC regional network of farm gate markets. From the government’s perspective, the experience also justified the role it had taken in 1998 to implement a national inputs subsidy programme that provided a ‘starter package’ (comprising fertilizer and maize seed) to all smallholder households, enabling them to enhance maize

Social protection, growth and employment

production. The downscaling of this programme in 2001, on donor advice that the measure was unaffordable, demonstrated the importance of input subsidies to smallholder output. However, the Mluzi government had become fiscally (and intellectually) too dependent on donors to refashion policy into a strategy that could build on these lessons. From about 1994, the rural population began to benefit (more systematically) from a range of ‘project’ interventions focused on improving livelihoods through wealth transfers and technical support. Although the scale of these interventions was ultimately modest, excluding the majority of households, their impact over the long term (up to the present) has probably been greater than often recognized. These interventions reflected a new recognition among donors of the need for social protection, not least because of the failure of structural adjustment to achieve growth that would benefit poor people. The World Bank, for example, funded a Poverty Alleviation Programme (PAP), commencing in 1994. PAP established a mechanism for wealth transfer through public works programmes (PWP) involving food and cash transfers. Although these measures were conceptualized as safety nets for ultra-poor and destitute people, the occurrence of large-scale shocks, such as droughts, highlighted the livelihoods vulnerability of even the nonpoor. These episodes underlined the need for agricultural productivity enhancement measures and extension support on a large scale. By about 2002, the government had elevated food security from a function of disaster preparedness to a central pillar of its agricultural policy. The new thinking on agricultural development (and marking the consensus that had been achieved on this topic between government and donors) was expressed in the World Bank-/IMF-funded Malawi Poverty Reduction Strategy (MPRS). A medium-term development strategy to leverage growth towards the Vision 2020 objectives and progress towards the Millennium Development Goals (MDGs), the MPRS comprised four strategic pillars. Under the pillar of sustainable pro-poor growth, the MPRS provided a safety net blueprint (National Safety Nets Programme) to improve the “quality of life of the most vulnerable”. The strategy had several elements, including the provision of free agricultural inputs (fertilizer and seed) to targeted smallholders with spare land capacity, but facing capital constraints. Furthermore, it recognized that “some sections of the population […] will need direct assistance” through providing “moderate support to the transient poor and substantial transfers to the chronically poor” (GoM, 2002: 64). The wa Muthrika era (2004–2012) President Mluzi was succeeded in 2004 by Bingu wa Mutharika. The new president was a former technocrat, brought into politics to lead the UDF. After one year in power, he left the UDF and established the Democratic Progressive Party (DPP). His main rallying cry in assembling political support for the DPP was the promise to eliminate corruption and the commitment to sustain fertilizer subsidies to smallholders. In leadership style, President wa Muthrika sought to emulate aspects of President Banda’s leadership, thus personally taking over supervision of the key Ministries of Agriculture and Education and reimposing discipline within the civil service. President wa Muthrika’s main (positive) contribution towards agricultural development was his determination to sustain input (fertilizer) subsidies. While this strategy draws from the experience of the

Malawi

Banda and Mluzi eras, wa Muthrika pursued this policy with a single-minded determination. His Agricultural Inputs Subsidy Programme (AISP) would be recognized, politically, for its contribution towards the achievement of record maize yields and above-average production for the period 2005–2010. This would enable President wa Mutharika to survive politically (in his first term) and consolidate his once tenuous hold on power. Later in this article, we analyse the impact and development implications of inputs subsidies. President wa Muthrika was initially successful in holding out against donor demands for (further) liberalization of agricultural markets and reforms to strategic parastatals. Apart from safeguarding the role of ADMARC, the NFRA and SFFRFM, the government’s agricultural policy built upon the MPRS foundations and sought to enhance the provision of agricultural extension and subsidies to small farmers. The policy direction was articulated in the 2004 Malawi Economic Growth Strategy (MEGS) and then taken forward in the 2006 Malawi Growth and Development Strategy (MGDS). There were, at this point, varying degrees of donor support for this direction. Although the donor community (including the World Bank) recognized that input subsidies (along with free inputs) had benefited small farmer agriculture growth, their major concern focused less on the mechanism itself than on its scale (and hence costs), the impact of parastatals on market development and the absence of evidence of welfare benefits. The evidence in 2006, as articulated in the MPVA, was that enhanced maize production had not reduced the depth of poverty or lessoned vulnerability. The MPVA report recommends that, if the government is to introduce subsidies, then these should be ‘market smart’, meaning that such interventions should enhance the development of private-sector fertilizer markets (GoM, 2006). The MGDS, which covers 2006–2011, outlines the current development strategies. These include measures to enhance agro-processing in the main export crops (tea, tobacco and sugar); measures to improve smallholder productivity through technology uptake, infrastructure development; measures to improve extension services (including training field advisory staff); and measures to ensure national food security. The latter include maintaining the strategic grain reserve, using public works programmes as safety net measures and enhancing government capacity to manage food aid interventions. The MGDS also includes a social protection strategy, intended to safeguard the most vulnerable groups from shocks and to strengthen livelihoods. The details of this strategy have since been clarified through the National Social Support Policy, issued in 2009. The policy sets out four major objectives of social support: first, to provide welfare to the “very poorest members of society” through predictable transfers; second, to protect the assets of poor and vulnerable households from depletion and erosion through PWP and social insurance schemes; third, to increase the productive capacity of poor and vulnerable households through a range of measures, including input subsidies, cash transfers, PWP, promoting village savings and microfinance programmes, and conditional cash transfers; and, fourth, to foster stronger linkages between ministries and to harmonize actions.

Social protection, growth and employment

Interventions and impact Since the early 1990s, social protection measures have fulfilled an important role in smoothing consumption for vulnerable households, especially those that have had to confront covariate shocks, such as food crisis. In this time, the range of social protection measures has evolved from classic short-term safety net measures to more long-term development-oriented actions. The main social protection mechanisms that have been used and their scope are summarized in Table 2.5; the number of beneficiaries and costs are no doubt an underestimation of the true scale. The World Bank (2007) undertook a study in 2006 to review the effectiveness of measures then in operation. The study sought to address questions of scale, impact at the household level and cost-effectiveness. It found that, over 2003–2006, annual project costs for social protection (including input subsidies) were about US$134 million (equivalent to about 15 percent of the annual government budget). The costs per beneficiary, it noted, varied significantly between different instruments. The least cost-effective interventions were relief cash, in-kind transfers, food transfers and cash for public works programmes. The most cost-effective interventions were unconditional cash transfers, input subsidies and input transfers. Later in the analysis, we consider the potential cost-effectiveness of food aid verses input transfers and input subsidies to address the ongoing challenge of missing food entitlements. Food aid The greatest expenditure on social protection (corresponding with the broadest historical impact) has been on the transfer of food, principally maize, to food-insecure households, be they either transient or chronically insecure. The World Food Programme (WFP) was required to provide food aid to smallholders in 1987–1989, 1990–1991, 1992–1993, and from 1994–1995 to the present. Between 1987 and 2001, the total volume of food aid (mainly cereals) supplied through donors for drought relief and targeted safety nets was 1,413,850 metric tons, at a cost to government and donors of US$174 million; the food component alone cost US$73 million (Charman, 2004: 9, 59). The 2001–2002 food crisis brought into question the high costs and non-sustainability of food aid, given the state of vulnerability within the rural population and weak response to market liberalization (see Levy, Barahona et al., 2004; Levy, 2005). Although the scale of this crisis was much politicized, it required a significant humanitarian intervention to stabilize household consumption.10 In the 12-month period from June 2002, the resulting Joint Emergency Food Aid Programme distributed 184,317 metric tons of maize, 19,331 metric tons of pulses, 23,770 metric tons of corn soya blend and 1,827 metric tons of vegetable oil to over 23 million beneficiaries and a further 1.5 million school children benefited from supplementary feeding.11 The total volume of maize imported over this period (commercial and food aid) has been estimated at 788,538 metric tons at a cost of approximately US$201 million (almost the value of the 2002 tobacco exports) (Charman, 2004: 64–65). The immense cost of the required food aid programme (and commercial imports, which compelled the government to borrow on the domestic market through issuing bonds) emboldened the case for providing a universal input transfer programme for smallholders.

Malawi

Table 2.5: Social Protection Interventions, Estimated Beneficiaries and Accumulated Costs, c. 1994–2010 Intervention

Target

Beneficiaries

(including multiple beneficiaries)

Cost, million US$

Food transfers (maize)

Food-insecure households, both transient and chronically insecure

26,233,958

174,740

Ultra-poor, female-headed households and households caring for orphans Food and cash transfers

Food-insecure households, both transient and chronically insecure

<5050

Ultra-poor, female-headed households and households caring for orphans Food for work

Transient food-insecure households, able to work

<92,292

<10

Cash for work

Transient food-insecure households, able to work

<1,587,004

<273

Targeted Nutrition Children under five Programme Expectant and lactating mothers

6% of rural children under five

The chronically ill School Feeding

Primary school children

Approx. 28% of school children

Relief transfers

The destitute and households unable to work 1,225

Cash transfers

Female-headed households; ultra-poor households

<300,000

Starter Pack/ Targeted Inputs Programme

All smallholder farm families; resourcepoor smallholders with land and able to farm

13,091,724

127

Input subsidies (AISP)

Resource-poor (and non-poor) smallholders with land and able to farm

<6,250,000

<542.8

Inputs transfers

Resource-poor smallholder with land and able to farm

<3,701,350

<49.5

Inputs for work

Transient food-insecure households with land, able to work

<238,857

<27.4

Source: 1994–2003: Charman (2004), 2003–2006 World Bank (2007), MASAF website.

The primary lesson from the 2001–2002 crisis was that it was more cost-effective to subsidize inputs for food production, on the basis of the Starter Pack experience, than to import maize and distribute food aid (see Levy et al., 2005). The argument in favour of productivity enhancement as an alternative (and sustainable) agricultural development pathway drew support from government, some of the donors and international non-governmental organizations

Social protection, growth and employment

(INGOs). But while there was agreement on the need for subsidization of inputs to stimulate smallholder production growth, there was no consensus on the scale of the investment required. The leading donors (including the World Bank) were at that time unwilling to entertain increased expenditure on subsidization and would, in fact, argue for downsizing of the subsequent Targeted Inputs Programme (TIP). In the 2001/2002 TIP, the number of household beneficiaries was reduced from about 1.4 million households to about 1 million, thus benefiting less than half of all farm families. Targeted transfers In policy discussion on social protection, there was consensus (post-2002) that long-term transfers were required for particular cohorts. The beneficiaries were to include female-headed households, households with a high dependency ratio, households caring for orphans and the destitute. For these groups, food (maize) would be provided through targeted interventions such as PWP (food for assets), supplementary feeding for school children and orphans, and therapeutic feeding programmes targeting children under five, expectant and lactating mothers and chronically ill persons. The establishment of the MVAC and enhanced institutional capacity to monitor household food security in specific regions across the country now enabled geographical targeting. This allowed the government and its partners to identify and categorize the degree of risks among subregions and to highlight geographic areas facing particular crises as a result of local shocks. The identification of households requiring support within these regions was then to be undertaken through a process of community-based targeting whereby the community members themselves were required to specify individuals deserving support on the basis of nationally agreed social and economic criteria (MVAC, 2005: 217). These have tended to give greater weight to social criteria (such as orphan-containing households) rather than subjective measures of food insecurity or income status (i.e. proxy means tests). Since the country has become nationally self-sufficient in maize (c. 2007â&#x20AC;&#x201C;2009), the case for generalized food distribution has diminished. The achievement needs to be understood within a particular historical context (within a period of good rainfall, high commodity prices, and increasing fertilizer use among small farmers). It does not mean that food crises will no longer affect the majority of smallholders. Safety net programmes are likely to be still required, not least for very poor households, but also for those undergoing transitory crises in specific localities where crops fail through flooding or erratic rainfall. In providing social protection in these situations, the government (along with donors and INGOs) recognizes that direct cash transfers or cash-for-work programmes can address the needs of food-insecure households with equal effectiveness and indeed cause the least disruption to markets and future output.12 Among local development practitioners, the effectiveness of food transfers in reaching the most food insecure has, for several years, come under question. There is concern that community-based targeting, despite its participatory character, is subject to manipulation and inclusion and exclusion errors. The MVAC study, for instance, reported that the â&#x20AC;&#x153;targeting of food aid to the poorest households appears weakâ&#x20AC;? (GoM, 2006: 218). The challenges of effective targeting in food aid are no less than those of alternative social protection modalities, such as cash transfer programmes.

Malawi

This has meant that programmes have had to invest considerable resources at the outset of any intervention in targeting, registration and identification costs. Public works programmes After the 2001/2002 crisis, the use of PWP proliferated, with these mechanisms seen as a costeffective safety net mechanism. A strategic paper prepared for the World Bank, for example, advocated PWP as a “highly suitable countercyclical intervention” for regional and self-targeting (Smith, 2001: 25). The bulk of subsequent PWP interventions, building upon the MASAF experience, used a minimum wage as a self-targeting mechanism and required the beneficiaries to perform specific tasks such as road clearing on the basis of national task rates.13 The approach has since been refined, aligning it with district development plans and conditional aspects incorporated with the aim of bridging consumption shortfalls. It is thought that, because of their brevity, these interventions have brought little improvement to (permanent) household income, though they have reduced the need for asset depletion. The average project was implemented within one year; some were extended for longer, but few have run for three years or longer (World Bank, 2007: 31). The IHS2 found that only 5 percent of households reported benefiting directly from PWP over 2001–2003 (NSO, 2005). In those localities where PWP were implemented, about one fifth of adult men and women reported having attained employment. The targeting of beneficiaries has been imperfect geographically and within communities. There is evidence that, in practice, PWP recruitment is done on a ‘first-come, first-served’ basis determined by knowledge and accessibility to the project opportunities (Chirwa et al., 2004). PWP have also been criticized as a food security strategy because the period of need usually corresponds with the rainy season, when conditions are unsuitable for performing infrastructure works and would in any case draw labour away from the agricultural sector (World Bank, 2007: 25). Cash transfers PWP are unable to address the requirements of the chronically poor and people unable to work. The 2008 census recorded a population of 498,122 disabled persons, of whom 133,000 had visual impairments and about 108,000 had lower limb disabilities. Many of these individuals—along with a portion of the population affected by communicable disease (840,000 adults living with HIV, of whom 290,000 are in need of antiretroviral treatment, and 42,447 persons affected by tuberculosis)14 — are unable to work. On this basis of this limited evidence, there could be as many as 300,000 households whose ability to work is severely constrained through illness or disability. These individuals and households need sustained wealth transfers to accumulate sufficient assets and to build capacity. Since about 2002, INGOs, along with donors, have begun to lobby for a shift from PWP to direct cash transfers in certain contexts, targeting ultra-poor people. A cash transfer pilot programme was initiated in Mchinji district (Southern Region) in 2001/2002 to test the practicality and effectiveness of this approach in Malawi. The pilot distributed three kinds of transfers (cash, vouchers and a commodities package) to targeted beneficiaries in 54 villages (Devereux, Baulch et al., 2006: 85). The project highlighted the influence that high food (maize) prices could have on the scale of benefit in situations were maize was in short supply (as happens seasonally), thus undermining their purchasing power to acquire

Social protection, growth and employment

maize on local markets. The apparent success of the intervention, nevertheless, has resulted in the use of cash transfers as short-term safety net measures to assist vulnerable households fulfil their food entitlement, given the availability of food (at affordable prices) as a consequence of input subsidies/transfers. Other experiences with cash transfers have confirmed the finding that these measures are most effective in reducing food insecurity where food markets are stable and if the transfer is linked to some form of agricultural productivity-enhancing measure, such as enabling the beneficiaries to procure inputs (see World Bank, 2007: 62; Ntata, 2010). In the case of the Dowa Emergency Cash Transfer programme, it was found that sizable multiplier effects extend to small enterprises (Davies and Davey, 2008; RHVP, 2007a). A further pilot in cash transfers was initiated in Mchinji in 2006 through a joint UNICEF–government project containing an element of conditionality. The transfer was increased with family size, from US$4 to US$13 per month, while an additional amount (stipend) was paid for children attending school, though this aspect is not monitored. The pilot has since been broadened in scope and scale and has become positioned as the ‘primary poverty-reduction tool’ targeting the country’s ‘most destitute households’. In 2009, the project was extended from the pilot stage to 23,651 households and 92,786 beneficiaries in seven districts. Although there were plans to scale up the intervention to target 300,000 households per year by 2010, this has not been implemented (Miller et al., 2010: 483 and www.unicef.org/infobycountry/malawi_56675.html). Early assessments of impact conclude that cash transfers provide an effective instrument to reduce seasonal shocks, raise consumption and lessen poverty. There is evidence, though preliminary, of associated gains in adult health, child education, labour availability, food security and asset accumulation (Miller, 2009). The positive food security result is also evident in the Cash and Food for Livelihoods Pilot project, an organizational collaboration between the WFP and INGOs. The project had three transfer mechanisms: food (50 kg of maize), cash (sufficient value to purchase 50 kg of maize) and a mixed food and cash transfer modality. An investigation of the project impact found that the cash transfer modality was the most cost-effective of the three mechanisms in terms of increasing household food security by 1 percent against a baseline value (Audsley et al., 2010). While the use of cash transfers has gained much organizational support, the start-up costs have tended to exceed the cost–benefit of traditional food transfers and PWP. The need to reduce targeting errors requires a substantial investment in beneficiary identification. There are also significant logistical challenges to disbursing cash in a country such as Malawi, where formal banking institutions are only operational in the cities and major district centres. Although measures such as mobile cash vending facilities have been introduced to overcome these barriers, the service remains inadequate (many beneficiaries have to walk a considerable distance to access these mobile banks) and costly relative to the size of transfer. Impact The impact of food and cash transfers has been modest in terms of generalized poverty reduction or in strengthening the resilience of households, but they have been fairly successful in providing a safety net to facilitate the recovery of households from covariate shocks and food

Malawi

insecurity crisis between 1994 and 2004. The MPVA report (GoM, 2006) concluded that “the sum total of these various safety net activities has not been sufficient to reduce the overall levels of poverty and vulnerability”. The minor changes that these initiatives have brought, it is argued, are attributed to the difficulty that social protection programmes have experienced in targeting (World Bank, 2007; Devereux, Baulch et al., 2006). Across the breadth of projection measures under review, the majority of programmes have been subject to inclusion and exclusion errors, benefiting non-poor and neglecting sections of ultra-poor people or those in most dire need of assistance. The MVPA concluded that targeting problems beset all social protection measures, resulting in the exclusion of about half the poor people and inclusion of about 40 percent of the non-poor (GoM, 2006: 230). Some programmes (including the recent cash transfer measures) have sought to reduce the scale of these leakages through strengthening their methodologies, introducing, for example, community-based, multi-stage, participating approaches. There is insufficient evidence to judge the effectiveness of these approaches. Yet the reliance on community judgement poses many socio-cultural challenges. Within communities there are local perceptions and interpretations of ‘need’ (i.e. poverty) that do not accord with programme intentions and the focus on ultra-poor people. As the Malawian scholar Blessings Chinsinga notes from his investigation of targeting in the TIP winter maize programme, “[T]he recurrent argument against targeting, encountered in this study [from the informants], is that it makes no sense to target when the welfare of most rural households is essentially the same” (2005: 297). An important counter to targeting is the strength of cultural practices of reciprocity through which households are required (and expected) to redistribute wealth/food (through direct transfers) and opportunity (access to land and employment) to marginalized households. These practices have ensured a degree of trickle-down to ultra-poor people, because and in spite of the benefit afforded to the non-poor. While targeting remains a concern to development practitioners, given the financial constraints on programme means, the consequences of inclusion errors are probably less significant than the impact on very poor people through exclusion. This logic was initially applied in the Starter Pack programme. The main weaknesses in the impact of social protection measures are not mistargeting, but their scale (too small and too brief) and crisis-oriented approach (providing safety nets). In contrast, measures to enhance small-farm productivity through input transfers and subsidies have had a wider social reach, in absolute terms and in targeting poor people. In addition, these measures have been more sustained in implementation over the past decade and, as such, have enabled the beneficiaries not only to recover from crisis situations, but also to accumulate assets and to build household resilience. We now examine in detail their impact.

Social protection, growth and employment

Raising small-farm productivity Input subsidies Smallholders were slow to adopt ‘green revolution’ technologies. In the mid-1980s, the government increased its investment in the efforts of research institutions to develop seed and planting material that would be more readily adopted. In the case of maize, research focused on higher-yielding maize varieties suitable to smallholders, meaning that the plants had to be openpollinated rather than hybrid. Numerous appropriate maize varieties were developed, though these were not widely used with most smallholders continuing to plant local varieties with the seed recycled annually. After the 1991/1992 and 1993/1994 droughts, the need to raise maize productivity became of paramount concern. The major agricultural development concern that then dominated strategic thinking was the issue of declining soil fertility. The solution, argued technical specialists, was the greater use of inorganic fertilizers (in concert with improved land husbandry practices). Research had proven that good returns from the application of nitrogen fertilizer could be expected (Rockefeller Foundation, 1998). But encouraging smallholders to invest in expensive inputs would require a new approach. Up until this point, fertilizer subsidies had only benefited the better-off small farmers, especially commodity producers; at the high point of SACA operations in the early 1990s, less than one quarter of smallholders (approximately 350,000 beneficiaries) had access to input loans. The Ministry of Agriculture set about the task of finding a solution for resource-poor farmers, with donor support, through establishing a Maize Productivity Task Force. Its role was to frame recommendations for the optimal use of fertilizers by smallholders, taking into consideration the country’s different agro-ecological zones. The emphasis on fertilizer use efficiency led the task force to consider various strategic options for achieving the overriding mandate (higher yields), but also to provide specific recommendations for smallholder conditions (Rockefeller Foundation, 1998). One option was the provision of an input package. The effectiveness of this option had been demonstrated over 1994–1996, when maize seeds and fertilizers (plus root and tuber cuttings) were distributed to about 800,000 households, thereby contributing towards the stabilization of maize output in the areas involved (Harrigan, 2008: 242). The task force also investigated and made recommendations on the viability of using ‘green’ technologies, including crop rotation, inter-planting and agro-forestry. The outcome of its research was a ‘best-bet’ strategy that could serve the various criteria for facilitating widespread fertilizer adoption in Malawi. At this time, there was little debate around the possible dependency on input transfers; the strategy was based on the assumption than improving productivity would raise on-farm income and reduce food insecurity, which was rightly considered to be an important first step towards poverty alleviation. Their strategy was to provide a small inputs package to all households, across the territory, to enable them to plant an additional 0.1 hectare of maize. The package was to contain hybrid seed and appropriate fertilizer. It was expected that this would enable the beneficiaries to produce sufficient extra maize to

Malawi

feed their families for one month, equivalent to an additional 280,000 metric tons to national output, and thus addressing, at the household level, the missing food entitlement between seasons. The cost of the initiative was estimated at US$18 per household, requiring a budget of US$32 million per annum. The Mluzi government immediately bought into the proposed strategy, in recognition of its dual political and economic potential, aligning with its pro-poor rhetoric. The country’s donors, on the other hand, were more cautious, recognizing the potential conflict between input transfers and efforts to rebuild rural credit markets and farmer finance institutions. Similarly, there was also concern that the measure would impact on private-sector input markets, discouraging growth (as a result of displacement) and firm investment in expanding distribution networks.15 The World Bank (at first) and the United States Agency for International Development (USAID) would not commit financially towards the concept; DFID (UK) agreed with the need to provide smallholders with subsidized inputs and would contribute funds, though it saw the strategy as an emergency measure (i.e. safety net) rather than as a long-term solution. The EU embraced the rationale, but was committed to the Agricultural Productivity Investment Programme (APIP), an initiative in partnership with the private sector that advanced fertilizer (on favourable loans) to creditworthy smallholders.16 Against the weight of donor opinion, the Mluzi government went ahead with the ‘best-bet’ proposal, providing ‘starter packs’ to 2.8 million smallholder beneficiaries in 1998 and 1999, effectively covering most farm households throughout the territory. The cost in 1999 was reduced to US$25.2 million, which was carried by the GoM (US$12.5 million), DFID (US$4.3 million) and the World Bank (US$7.2 million) (Harrigan, 2008: 245). The Starter Pack programme had a dramatic positive impact on national maize output and household food availability. Despite this outcome and the case for sustained input transfers, the government remained dependent on donor support. Its revenue base was under pressure, partially as consequence of currency devaluation and partly from rising debt. So, in 2000, it heeded donor advice and reduced the number of beneficiaries by half to 1,430,000 households. The government contribution towards the costs (and thus its influence on the programme) began to progressively decline with the costs carried by DFID, the EU and the Norwegian Government (NORAD). The programme was renamed ‘TIP’ and the inputs package was amended, with openpollinated maize seed varieties (OPVs) replacing the hybrid seed and the quantity of fertilizer reduced from 15 kg per pack to 10 kg. In the 2001 version, the number of beneficiaries was further reduced to 998,499 households, while a legume crop (soya) was included in the packs. The beneficiaries of TIP were identified through a community poverty-targeting approach that relied heavily on the administrative support of village-level traditional leadership structures. It is unsurprising that targeting was problematic, resulting in inclusion (non-poor) and exclusion (deserving recipients) errors and causing acrimony between beneficiaries and non-beneficiaries. Although the intention was to refocus on very poor, including female-headed, households, almost 40 percent of the beneficiaries were actually non-poor. The IHS2 reported that the portion of household in the fifth wealth quintile (i.e. non-poor) benefiting from TIP rose from 23 percent to 28 percent between the 2001 and 2004 versions, while not more than 50 percent of households in the first (i.e. ultra-poor) quintile were targeted (NSO, 2005: 102). In food security terms, the impact that TIP had on household production was modest, delivering a net increase of 35 kg of

Social protection, growth and employment

maize per household (Charman, 2004: 52). The programmeâ&#x20AC;&#x2122;s limitations were soon exposed in the 2001/2002 food crisis, with critics attributing the poor maize harvest to the dramatic reduction in the scale of programme. After 2001/2002, the TIP continued for two more years. Its impact on productivity, however, steadily diminished. The donor community were beginning to turn away from this strategy, arguing that the programme was an unsuitable strategy for raising national maize production, given its targeting of resource-poor farmers who (it was thought) had little potential to increase agricultural output because of their land and labour shortage. Some donors began to argue for alternative (non-productivity) forms of social protection for very poor people, including sustained food aid, supplementary feeding and wealth transfers. With respect to the situation of smallholders, a new wave of programme thinking had entered donor circles that held that the solution to agricultural development required greater attention to institutional development17 (including farmer organizations and promoting the capacity of private service providers) and strengthening markets and building agro-industrial linkages. It was thought that the market (and not the state, via input transfers) should drive the process of modernization on farms. There was (justifiable) concern about smallholder dependency on free inputs (farmers were delaying their input purchases, awaiting hand-outs), while the absence of an evidential poverty reduction impact was worrisome to funders. Malawian NGOs and INGOs sought to sustain the case for fertilizer subsidies, based on their field experience of the success of Starter Pack/TIP,18 though their political leverage had begun to wane as food security recovery programmes (which enhanced their power) began to tail off. The wavering of donor support for input transfers happened at a time of considerable political debate on agricultural policy between the main contending parties in the run-up to the 2004 election. The main political contest was between a revived MCP led by John Tembo, a former aide to President Banda and doyen of the indigenous tobacco estate class, and the UDF, then under the leadership of Bingu wa Muthrika. Both parties sought to stake their claim as pro-smallholder through their promise to provide subsidized inputs universally (Chinsinga, 2008). After the 2004 election, President wa Muthirka was under pressure on two fronts: from donors, who demanded improved fiscal discipline if Malawi was to attain debt relief, and from the opposition bench, which demanded that he deliver on his word to support inputs subsidies. The opposition dominated the Parliamentary Committee on Agriculture and called for a universal fertilizer subsidy. The government made the bold decision to go alone into an Extended TIP (ETIP), providing a more comprehensive input package (26 kg of fertilizer and 5 kg of maize seed) to 2.8 million beneficiaries, almost all of whom were once again farm families. The inputs were to be provided through a voucher system. As there was no donor stake in this programme, the government decided to manage the importation of fertilizer and its distribution through parastatal entities, excluding the private sector while forcefully demonstrating its commitment towards reviving the institutional framework of the Banda era. As the government had made its decision to finance ETIP fairly late, there was insufficient time to import the inputs and to manage their distribution before the commencement of the new agricultural season. The late delivery of the packages and an unfavourable rain season resulted in relatively poor maize yields, though the

Malawi

packages themselves gave President wa Muthrika a strong political dividend through proving that he was committed to the policy of input subsidization. The dip in maize production in 2004–2005, which necessitated further food aid, again posed the question of whether production enhancement was not a more cost-effective solution than procurement from international markets. It was generally recognized that the poor maize harvests could be attributed to the operational setbacks of the programme (plus unfavourable weather), rather than to the measure itself. As the new government had begun to demonstrate its seriousness to maintain macroeconomic stability, donors were again willing to finance input subsidies (recognizing the potential cost-effectiveness of productivity enhancement over aid), though on terms that ensured that private-sector inputs suppliers and distributors would be afforded a role. The president remained under significant domestic political pressure, with the opposition advocating a universal inputs subsidy for maize and tobacco. Their demands were largely met through the introduction in the 2005/2006 season of the Agricultural (Farm) Inputs Subsidy Programme (AISP). AISP did not amount to a universal subsidy, but instead provided the majority of smallholders with vouchers that entitled them to purchase two 50 kg bags of fertilizer at a significant discount — in other words, the programme provided a price subsidy. It so happened that the introduction of AISP coincided with a particularly favourable season, with rainfall distributed evenly across the country throughout the season. As a result, there was a record-high maize yield. Food security was attained at the national and household levels. The productivity gain stabilized the inter-seasonal maize price, thus enabling the food-insecure to acquire maize at affordable prices, with national reserves rising to a level sufficient to permit exports. The initial success of AISP gave Malawian political interest groups—the government and the opposition—a policy victory over donor conservatism. The breadth of cross-party political support for subsidies has ensured that this strategy has continued to the present. Donors have since provided finance towards the programme, strengthening their influence over its design and funding monitoring and impact assessments. The original model entailed the distribution of vouchers (coupons) to farmers, subsidizing the input costs for maize and tobacco. The vouchers were issued to farmers via traditional leadership structures and then were redeemable at ADMARC/SFFRFM depots (and subsequently at private vendors). The value of the voucher was equal to about two thirds of the market price, a very sizeable discount. The maize subsidy covered 50 kg of 23:21:0:4 (N-P-K-S) (plus maize seed), whereas the tobacco voucher was for 50 kg of calcium ammonium nitrate and 50 kg of D compound (8N-14P-7K) (Dorward et al., 2009: 15). The model has been refined over the past four seasons: the inputs package, distribution and redemption system, for instance, have all been amended. These changes enabled farmers to use the package for a wider range of crops (including legumes, cotton, tea and coffee, whist beneficiaries could choose between hybrid maize or OPV seed) through a wider choice of vendors, including private companies. The approach to targeting has been revised, supposedly to provide a fairer allocation per farm household, using data from a central register maintained by the MoAFS. For the 2009/2010 season, the government has had to scale back the programme due to its inability to sustain the high costs. In this season (2009/2010), the MoAFS will provide vouchers only for maize production, thus no longer providing subsidized inputs for tobacco

Social protection, growth and employment

and other cash crops (Holden and Lunduka, 2010b: 33). The expectation is that smallholder commodity producers (along with the tobacco estate class, who had originally led the political demands for these subsidies) will procure inputs for non-maize crops at market rates, using their savings and increasing household asset position. The costs of AISP escalated from US$51.4 million in 2005/2006 to US$284.6 million in 2008/2009, reflecting the escalation in the programme size and implementation costs; the latter figure equates to about 74 percent of the total MoAFS budget (Dorward et al., 2010: 22). Experts consider that its main design weaknesses were the difficulty of selecting beneficiaries and the logistical complexity of procuring, distributing and validating the vouchers. The vouchers themselves present an opportunity for theft and fraud, representing a value of 10 percent of the annual household income of about 40 percent of the population regarded as poor. There is evidence that some of the beneficiaries resold their vouchers (thought the extent of this practice was relatively small) and benefited through income rather than increased productivity. It may be argued that the informal resale of coupons was correcting mistargeting, with the transfer value thus divided between the (poor) seller and the (non-poor) buyer. Some of the initial criticisms of the voucher system have been overcome through experience and refinements to the approach, though the role of traditional leaders in selecting beneficiaries remains an obstacle to transparency and accountability. There are objections that the programme has not benefited ultra-poor people (many of whom are unable to meet the unsubsidized portion of the input costs), although, over this period, the position of these households has improved, suggesting a strong trickle-down effect. Impact on crop production The government claims that AISP is a success. In terms of maize production (output) and productivity (return on investment), there is evidence to support this claim. While these achievements have resulted in much political backslapping, there are legitimate academic concerns about the validity of the national maize yields (with commentators accusing the government of ‘sexing up’ the results).19 But it should be noted that the manipulation of crop statistics is not new and has probably taken place for much of the past two decades, sometimes upwards and at other times downwards, to serve different agendas. This will remain a challenge into the future so long as accurate data is difficult to acquire, given the methodological complexities of measuring non-marketable crops, grown under smallholder farming systems. We can nevertheless be fairly certain that AISP resulted in bumper maize harvests and above-average yields in several crops from 2006/2007 through to 2009/2010. We now consider the evidence from maize, tobacco, and roots (cassava), tubers (sweet potatoes) and legumes. Maize Maize is symbolic of national agricultural ambitions and food security. The importance of maize is acknowledged within the iconography of the MoAFS through the incorporation within its logo of a maize plant and maize granary. The government (along with donors) articulates national food requirements in terms of maize. This has been described as the “politics and mass patronage of maize self-sufficiency” (Dorward et al., 2008). The importance of maize as a food staple for

Malawi

smallholders was quantified in the IHS2. The survey found that 97 percent of rural households grow maize; so, too, do nearly all urban households through rural gardens/fields cultivated by tenants or relatives (NSO, 2005: 95). The aim of achieving household maize self-sufficiency has become internalized as a desirable objective among smallholders across the territory, regardless of wealth or geographic location (and despite agro-ecological suitability to produce maize). At present, the national maize requirement to satisfy Malawiâ&#x20AC;&#x2122;s food security needs is approximately 2.5 million metric tons (FEWNET, May 2010). Over the past five seasons, smallholder output has clearly exceeded these requirements. Though (as indicated above) the precise scale of this achievement is uncertain, there is no doubt that a significant output increase has occurred since 2005. The production trends in maize (distinguishing between seed varietals: local, composite or OPV, and hybrid) from 1987 to the present are shown in Figure 2.2. It should be noted that the significant growth in maize output from 2005 correlates with a decline in poverty (over the same period) and, given that maize is the main smallholder crop, the attribution is strong. In the decade from 1987, smallholder maize production began to falter, reaching its lowest point in 1991/1992 with a total output of 657,000 metric tons. The main causes of this decline

Figure 2.2: Maize Production, Disaggregated by Variety, 1987â&#x20AC;&#x201C;2010

Source: MoAFS Final Crop Estimates

Social protection, growth and employment

are thought to have been degrading soil fertility through poor husbandry practices, low input uptake and the limited scale of green technology adoption. In 1991/1992, the influence of these factors was compounded by low rainfall. As a response to that crisis, the government (through World Bank funding) distributed fertilizer and hybrid maize seeds to all smallholders. Production swiftly recovered in 1992/1993, with approximately half of the total output grown from hybrid seed. The resilience of smallholders remained weak and, against the backdrop of the political turmoil of the transition process in 1993/1994 and drought, production again collapsed. It took four seasons before the pro-poor agenda of the new government was translated into a universal maize productivity enhancement programme. In the intervening period, a combination of free input programmes, the liberalization of the seed market and technology promotion20 helped to sustain the shift from traditional seed technologies (seeds with no research input) to higheryielding varieties (derived from research output), whose value to smallholders had now been demonstrated. Under the conditions of declining soil fertilityâ&#x20AC;&#x201D;a consequence of poor agronomy and small landholdingsâ&#x20AC;&#x201D;productivity gains could only be achieved through the application of fertilizer inputs, especially nitrogenous compounds. The finally achieved enforced removal of fertilizer subsidies in 1996/1997 (along with currency devaluation) resulted in a dramatic decline in smallholder consumption of Urea, with national consumption falling from 38,167 metric tons in 1995/1996 to 1438 metric tons in 1996/1997. This experience provided the evidential foundation for the introduction of a fertilizer subsidy for non-poor people along with a free inputs transfer programme targeting poor people. The latter came into effect in 1998/1999. The replacement of the Starter Pack programme with the TIP saw maize yields decline. Total national maize production (also) fell considerably after 1999/2000 and continued along a downward curve over the course of the five subsequent seasons. The decline has been attributed to the reduction in fertilizer uptake and market influences. The exclusion of nonTable 2.6: Maize Fertilizer Consumption, poor people from the TIP undoubtedly 1996â&#x20AC;&#x201C;2007 (metric tons) affected output. There is evidence (in the Year /Fertilizer Urea 23:21:0:4S historical consumption of fertilizer and in tobacco data) that non-poor people chose to rationalize their resources by focusing on tobacco over maize. Smallholder uptake of fertilizers for maize (urea and 23:21:0:4S) rose, though unevenly, throughout this period (as shown in Table 2.6), with uptake rising fastest from 2002/2003 onwards. Input transfers/subsidies evidently did encourage greater use of inorganic fertilizer and higher-yielding seeds. This is most noticeable in the shift from local maize seeds towards higher-yielding varieties,

1996/97

1438

27189

1997/98

39450

35145

1998/99

42116

33188

1999/00

41150

41120

2000/01

35044

36713

2001/02

36719

38467

2002/03

60539

52478

2003/04

57481

53232

2004/05

71556

78085

2005/06

99442

96291

2006/07

115505

102682

Source: RBM (2009)

Malawi

hybrids and OPVs alike. The inclusion of OPVs in the inputs package gave smallholders an experience of the potential benefits of higher-yielding technologies, with these composite seeds able to produce almost double the yield of local varieties. Since this time, OPVs have progressively displaced indigenous low-yielding varieties, which now account for less than one third of output, though they are still popular for the palatability of the maize. AISP had an exponential effect on maize production. The dramatic reduction in input costs made maize more competitive with marketable crops, encouraging farmers to produce surplus. The programme significantly increased the total level of fertilizer consumption, even though it displaced sales through private vendors. The scale of this displacement effect was between 30 percent and 40 percent in 2006/2007 (Dorward et al., 2008), prior to the introduction of programme changes that have strengthened the role of the private sector. From a low base point, the volume of urea consumption rose to 115,505 metric tons by 2006/2007 (latest data), while the consumption of 23:21:0:4S rose to 102,682 metric tons. In the case of the latter input, smallholder consumption grew 377 percent over the course of the decade after 1994. Although AISP included non-poor people within its target, the impressive growth in maize production owes much to the endeavours of poor people. Although the non-poor had a role in increasing production, the scale of their investment in maize was constrained by their simultaneous expansion of cash crop production, notably burley tobacco. AISP originally provided subsidized inputs for these crops. Most smallholders in the higher consumption quintiles had access to neither unused land nor surplus labour to expand maize production; in most cases, these resources (land and labour) were stretched in the pursuit of cash crop opportunities. Instead, the main drivers of the increase in maize production were poor smallholders, i.e. those households with available land and surplus labour, who used the inputs subsidy to increase the area under cultivation while enhancing their productivity through using higher-yielding seeds and fertilizer. AISP enabled these producers to achieve increased household maize production with less labour, which, in turn, permitted them to sustain their efforts in crop diversification and in the pursuit of off-farm income opportunities. It is noteworthy that the increase in maize production in this period was achieved on roughly the same size land area as had been planted to maize at the start of the decade. However, among the land-poor, research has shown that input subsidies (and ASIP, in particular) have enabled smallholders to reduce the area allocated to maize (currently an average of 0.71 hectares on 1.17-hectare farms), thus effectively releasing land for other crops and allowing for improved soil management practices (see Holden and Lunduka, 2010a). In the 2008/2009 season, about 65 percent of farm households received one or more fertilizer coupons. The targeting process was said to be biased towards households that were engaged full-time in agriculture and had more land and more assets, thus reducing the proportion of benefiting households among the ultra-poor (Dorward et al., 2009: 20). The marginalization of poor people, especially those with land and labour, is a strategic error, given the higher productivity of small farms (see Conclusion).

Social protection, growth and employment

Burley tobacco The Banda government excluded smallholders from growing tobacco; this decision was made to preserve opportunities in the tobacco sector for the estates owners, who thus needed (cheap) labour. It has been argued that the MCP rulers did not want competition from smallholders and would not have been able to alienate land and secure labour, had smallholders been afforded the same opportunity (see Weyl, 1991). Tobacco production requires a substantial investment in fertilizers and sustained labour input for nine months. Under SAL pressure, the government initiated a smallholder burley (air-cured) tobacco scheme in 1990 whereby a small number of smallholders were permitted to grow tobacco under specific conditions. The scheme was gradually expanded, though the relatively low prices on offer through ADMARC (which monopolized market access) dampened the smallholder response. After 1994, two developments would provide new opportunities for smallholders to produce burley. First, the establishment of the Malawi Rural Finance Company provided a mechanism through which smallholders could obtain loans for inputs issued through group structures. Second, the repeal of the Special Crops Act (de facto in 1995) meant that all smallholders, regardless of their situation, could not only produce tobacco, but also market their crops on the auction floor. Smallholders responded to these opportunities in large numbers, raising burley output significantly; the trend and yield is illustrated in Figure 2.3.

Figure 2.3: Burley Production and Yield, 1987â&#x20AC;&#x201C;2010

Source: MoAFS Final Crop Estimates.

2009/2010

2008/2009

2007/2008

2006/2007

2005/2006

2004/2005

2003/2004

2002/2003

2001/2002

2000/2001

1999/2000

1998/1999

1997/1998

1996/1997

1995/1996

1994/1995

1993/1994

1992/1993

1991/1992

1990/1991

1989/1990

1988/1989

1987/1988

Production (â&#x20AC;&#x2DC;000 KG)

burley production burley yield

Malawi

From the late 1990s, smallholder production began to exceed estate production. The number of tobacco growers and producer clubs rapidly multiplied, aided in part through the support of a farmer organizational development programme. By 2004, there were approximately 18,288 registered tobacco clubs, representing more than 300,000 individual smallholders (Koester et al., 2004: 88). Since this time, the number of individual growers has continued to multiply, with a significant portion now engaged as ‘freelance’ producers neither belonging to small farmer clubs nor individually registered with the tobacco authorities; many of these smallholders fall within the socio-economic category of poor. The provision of AISP subsidies to tobacco growers spurred growth in smallholder burley output, resulting in a sharp improvement in tobacco yields and hence in productivity. Total output increased from 2006/2007 to 2008/2009, rising from 107,309 million kg to 193,911 million kg. It is important to note that the smallholder response was strongly influenced by favourable (i.e. higher) auction floor prices, showing that smallholders have the capacity to raise output when incentives are strong and costs low. Since the input subsidies for tobacco were ended (also taking into consideration higher labour costs), the price incentive has diminished. The reduction in the AISP package for commodity producers since 2008/2009 has evidently impacted yields and output, though the trend is too short to draw a firm conclusion on the impact. Roots, tubers and legumes In maize and tobacco, smallholder output is inextricably tied to fertilizer distribution and/or subsidization. In other food crops, such as roots, tubers and legumes, smallholders do not use inorganic fertilizers, while seeds/planting materially are usually recycled. In terms of agricultural research, far less effort has (historically) been directed towards these crops, especially in terms of developing improved varietals. Since about 1994, smallholder output of cassava, sweet potatoes and legumes (ground beans, peas and soya) has increased substantially; the trend is shown in Figure 2.4. Roots, tubers and legumes are important staples in poor households, cultivated by women for the main purpose of meeting household food requirements. The IHS2 reports that 24 percent of rural households grew cassava (45 percent in the Northern, 25 percent in the Southern and 12 percent in the Central Regions, respectively) and 51 percent grew pulses or legumes (comparatively even distribution) (NSO, 2005: 98). Over time, more and more households have cultivated cassava. The MVPA reports that some of the households who grew cassava in 2005 did so as a ‘consumption-smoothing’ mechanism to counter maize shortage (GoM, 2006). Yet, despite the increase in national maize availability, the trend in cassava and sweet potato production has remained on a (steep) upward curve. It is important to note that these two crops are not grown in significant quantities by non-poor people, whose efforts and resources are (wholly) tied up in the production of maize and tobacco. The rise in the production of cassava, now Malawi’s second most important food staple, has been attributed to drought and labour shortage because the crop has much lower water and labour requirements. But these explanations are inadequate. They cannot account for the expansion of cassava into higher rainfall areas and the increased net allocation of household labour

Social protection, growth and employment

Figure 2.4: Cassava and Sweet Potato Production, 1987â&#x20AC;&#x201C;2010

2009/2010

2008/2009

2007/2008

2006/2007

2005/2006

2004/2005

2003/2004

2002/2003

2001/2002

2000/2001

1999/2000

1998/1999

1997/1998

1996/1997

1995/1996

1994/1995

1993/1994

1992/1993

1991/1992

1990/1991

1989/1990

1988/1989

1987/1988

Production MT

cassava sweet potatoes

Source: MoAFS Final Crop Estimates.

afforded to it and to other traditional crops. Instead, the increased investment of labour towards cassava, sweet potatoes and legumes may be the result of the labour savings and the release of land from maize cultivation through the productivity gains arising from inputs transfers/ subsidies. The data shows a close correlation between production spikes in cassava and sweet potatoes and the universal provision of fertilizer and maize seed inputs to all smallholders. In socio-cultural terms, the main beneficiaries of this multiplier effect were households in general, who enjoyed more food security, and women in particular, who gained entitlements or control over these traditional crops. Since the mid-1990s, the production of traditional food crops has been strongly promoted in multiple INGO projects and within social protection measures. For example, cassava and sweet potato cuttings were distributed in drought recovery interventions, through direct transfers and PWP; similarly, legume seed was included in input transfer programmes. The success of these initiatives, it has been argued, owes less to the availability of the planting material/seed than to the targeting of these interventions to marginalized (i.e. ultra-poor) households and to women in particular (Charman, 2008). In cultural terms, roots, tubers and legumes are crops in which female entitlements (or use rights) are relatively secure, meaning that women could retain ownership of the resource and legitimately devote land and labour towards their production in accordance with customary practice. These crops are culturally regarded as â&#x20AC;&#x2DC;belongingâ&#x20AC;&#x2122;

Malawi

to women in the sense that they have entitlement to use them to feed their families or to sell them for income, as they choose appropriate. The promotion of cassava, sweet potatoes and legumes through social protection interventions under the banner of a government-sanctioned food security strategy reinforced their control over these crops.21 Hybrid maize seed and fertilizer, in contrast, have historically been resources in which men have held dominant entitlements through their more advantaged position to acquire finance and their customary ownership of the household (maize) granary. The MPVA report (GoM, 2006: XX) comments on this dynamic: “While women hold decision making power in female-headed households, in male headed households there is a clear division: to the extent that women are involved in decisions about inputs and planting, their role is largely limited to crops that do not require fertilizer application, and where seeds are recycled.” The simultaneous provision of hybrid/OPV maize and fertilizer in the inputs transfer programmes may have detracted male concern from the strengthened entitlements of women in traditional crops, permitting them time (labour) and land to intensify production without opposition or significant contestation. Cost–benefit considerations The Starter Pack and TIP programmes (input transfers) were shown to transfer benefits (in terms of increased household maize production) more cost-effectively than procuring and distributing grain through food aid programmes. An evaluation of the 1999/2000 edition estimated that the maize productivity gain arising from the US$20 million programme (280,000–420,000 metric tons) would have cost between US$70 million and US$100 million, had the maize been obtained through commercial or food aid pipelines (Levy, 2005). The evaluation considered the major known cost variables (fertilizer procurement, logistical and operational costs) and then examined the qualitative impact of the programme on households through surveys.22 While cost-effective in comparison to food aid (see below) and certainly of benefit to the smallholder beneficiaries, input transfers on the scale undertaken would prove to be inadequate as a means of substantially increasing national production and stabilizing supply conditions. Most of the additional maize did not enter markets and was consumed at home. The measure was incapable of redressing crisis situations where households needed to recover from falling into poverty, whereas the input transfers could provide no more than a minor productivity boost. A more appropriate pathway, from the government perspective, was the provision of input subsidies on a scale that would enable farmers to substantially raise production, returning to the development concept first used under the Banda administration but now broadened in scope to include poor farmers. There is no doubt that AISP achieved this goal. The cost–benefit of the programme is hard to determine because of the complexity of the numerous variables that influence the productivity outcome. An in-depth assessment of the 2006/2007 programme found that the benefits (a gain of between 423,000 and 881,000 metric tons) needed to be weighed against crude programme costs (US$52.3 per beneficiary) and effectiveness of the measure (SOAS et al., 2008). The analysis drew on qualitative data from focus group discussions with farmer groups in 12 districts, key informant interviews with programme stakeholders and beneficiaries, and an examination of descriptive statistics. The cost–benefit analysis used econometric modelling and livelihoods modelling. In its assessment of costs, the study considered: i) the displacement costs to the

Social protection, growth and employment

commercial fertilizer market (which varied between 30 percent and 50 percent); ii) the programme operational costs; and iii) the incremental gain from the use of nitrogenous fertilizers, a variable that is itself subject to rainfall conditions and crop management practices. In terms of the overall objective of increasing input availability, the ASIP 2006/2007 programme cost US$1.19 to transfer US$1 of fertilizer, which, in turn, produced a productivity gain in maize of between 5 kg (low level) and 11 kg (upper level). This cost benefit was evidently much lower than the input transfer programmes that, although their returns are surely exaggerated through the exclusion of such variables, tempted the authors to call for a reduction in the programme scale and size of input packages. AISP costs have escalated in the subsequent two seasons (rising from US$91 million to US$284 million) (through a combination of factors, including the increase in programme sale, rising fertilizer costs and inefficiencies in implementation), while the gain in productivity indicates a possible diminishing return (although no detailed analysis has been conducted). On the basis of the evidence from the Starter Pack programme and AISP, we can contrast the cost benefit of these measures against that of direct food transfers. In the 2009/2010 crop season, as already mentioned, the MVAC identified a missing food entitlement for 275,000 persons over five months. The report motivated for the provision of food aid (in terms of 50 kg of maize) on a monthly basis to the vulnerable 68,750 households. Using data on the costs of a WFP social protection programme in Malawi that was undertaken at about this time, we can estimate that the cost of procuring and distributing the necessary food (through a food aid pipeline) would amount to US$7.4 million. This equals an investment of US$109 over the five-month period to mitigate hunger in one household. The alternative cost of supporting these households to produce sufficient food to fulfil the missing entitlement, were this option possible prior to the onset of the crisis situation, is then examined. The results, as shown in Table 2.7, indicate that the Starter Pack approach is the most cost-effective, requiring a total investment of somewhere between US$800,000 and US$1.2 million, or between US$12 and US$18 per household, to produce sufficient maize to survive this period. In contrast, the AISP subsidy would have cost considerably more, requiring a total investment of between US$1.5 million and US$3.4 million, though the measure is probably more likely to have supported market conditions (stabilizing maize supply and smoothing price spikes). Our analysis shows that the cost of enabling one household to produce sufficient maize to fulfil this missing food entitlement over five months through this modality would range between US$23 and US$50 per household, subject to the variable of incremental productivity gain from fertilizer adoption. These costs are still more than 50 percent lower than those of food aid. Despite the evidence in their favour, there is concern that input transfers and subsidies have been unduly costly (and inefficient) and are not sustainable, given the budgetary constraints on the GoM. The 2006/2007 ASIP assessment concluded that the costâ&#x20AC;&#x201C;benefit ratio ranged from 0.76 to 1.36, implying that the measure could indeed be economically justifiable if the programme was well managed (SOAS, 2008: IV). But its affordability remains doubtful and can be implemented on a significantly wide scale (i.e. universal distribution) only with donor funding at a cost of about US$200 million per year. Yet sustained donor commitment is unlikely, not least because opponents of subsidization perceive other measures (including food aid) as more

Malawi

Table 2.7: Cost–Benefit Comparative Analysis: Starter Pack, AISP and Food Aid Social Protection Measure Evidence

Starter Pack

Year Fix cost Beneficiaries

ASIP

Food Aid

1999/2000

2006/2007

USD 20,000,000

USD 91,577,250

2800000 HH

1750000 HH

USD 7.1

USD 52.3

Cost per beneficiary

2008/2009 WFP Pilot Experience

Maize price (FOB) (ton)

USD 306

Distribution and handling costs (ton)

USD 120

Transfer costs (per USD 1 expended on fertilizer)

USD 1.19

Food transfer (per USD 1 expended on maize)

2.3 kg maize

Maize productivity gain (upper level)

150 kg/ha

881000 mt

Maize productivity gain (mid level)

100 kg/ha

413000 mt

Productivity return on USD 1 transfer (upper level)

21 kg maize

11 kg maize

Productivity return on USD 1 transfer (mid level)

14 kg maize

5 kg maize

Cost / Benefit in current situation 2009/10 Food-Insecure Population

275,000

68,750

250

17,187,500

Total cost of mitigating hunger for 1 HH (upper productivity level)

USD 12

USD 23

USD 109

Total cost of mitigating hunger for 1 HH (lower productivity level)

USD 18

USD 50

Total cost of mitigating hunger (upper level)

USD 818,452

USD 1,562,500

Total cost of mitigating hunger (lower level)

USD 1,227,679

USD 3,437,500

Households in need Missing food entitlement (kg) per household: 5 months @ 50 kg maize Total missing food entitlement (kg)

Source: Levy, 2005; SOAS et al., 2008; Audsley et al., 2010

USD 7,472,826

Social protection, growth and employment

Social Protection Measure Variables

Starter Pack

ASIP

Food Aid

Effectiveness of targeting (minimizing inclusion and exclusion errors) Programme logistical efficiency, including costs of targeting & registering beneficiaries and monitoring impact. Incremental grain production from use of nitrogenous fertilizer subject to horticultural / management practice as well as rainfall. Inflexible tool for dealing with national food shortage and market inefficiencies as increased production will not necessarily become available as marketable surplus during demand peaks.

Effectiveness of targeting (minimizing inclusion and exclusion errors) Programme logistical efficiency, especially role of parastatals whose relatively higher costs impede market competitiveness. Incremental grain production from use of nitrogenous fertilizer subject to horticultural/ management practice as well as rainfall. Market displacement, i.e., decreasing commercial sales (between 30% and 50%). Inflexible tool for dealing with national food shortage and market inefficiencies as increased production will not necessarily become available as marketable surplus during demand peaks.

Effectiveness of targeting (minimizing inclusion and exclusion errors) Food pipeline breakages or disruptions. Transfer value is subject to considerable fluctuation, especially when grain procured on international markets is below seasonal spikes within the domestic market (April-August): the price gap provides an operational subsidy. Effective tool for humanitarian crisis situations, though costs are not necessarily lower than cash transfers. The benefit for agricultural development is limited.

Source: Levy, 2005; SOAS et al., 2008; Audsley et al., 2010

cost-beneficial (ignoring the significant costs of targeting and leakages), partly because the food component is seen as a benefit rather than as a cost23 and partly because food aid (along with other forms of transfers) is a more flexible instrument that can be mobilized in short lead times to address production shortfall shocks. The argument in favour of input transfers remains strong within government policy decision-making circles, reflecting the above evidence. There is still significant room for improving the programmeâ&#x20AC;&#x2122;s efficiency and rationalizing (or downsizing) the input package to ensure maximum possible return on investment. The 2006/2007 assessment indeed advocates comprehensive coverage of all smallholders with a reduced subsidy (SOAS et al., 2008: 100), thus keeping costs down while ensuring near universal coverage and thereby minimizing mistargeting costs.

Impact on labour markets Free inputs, input subsidies, food/cash transfers and PWP have positively and negatively affected rural labour markets. As discussed earlier, some of these programmes have affected labour market participation through offering above-equilibrium wage rates. PWP have been implemented in about 14 percent of communities (GoM, 2006: 134), though participation on average is usually brief (for a few weeks) and only 2 percent of households reported that food-for-work (FFW)

Malawi

provided their main source of food (NSO, 2008: 100). A more significant impact on rural labour markets has resulted from the decline in surplus labour during the height of the agricultural season consequent to greater on-farm productivity. Across the country, rapid wage increases were observed after the 2006/2007 season. The AISP evaluation (SOAS et al., 2008: 33) found that the improvement in maize production and enhanced household access to food (maize, but also cassava) reduced the need of poor people to ‘hire out’ labour to meet short-term cash or food needs. At the same time, the availability of subsidized inputs increased the ability of non-poor smallholders to ‘hire in’ labour to enhance their agricultural productivity through expanding their fields under cultivation and allocating additional labour to crop management. These labour market pressures increased piece-work rates, which translated into a positive income effect for households hiring out labour. The increase in piece-work wages was observable throughout the country, though most pronounced in areas of greater labour scarcity and lower relative poverty. The scale of the rate increase varied from about 33 percent for brief tasks such as weeding to two- and three-fold increases in month-long work assignments. The authors of the investigation postulate that the combination of falling maize prices (due to increased maize supply) and rising piece-work rates could significantly reduce affect the real income of those households hiring out their labour to earn food. There have been similar wage spikes in communities benefiting from cash transfers where the target group is those households that would otherwise need to seek piece-work to secure food. These wage increases are characteristically short-lived and tend to return to an equilibrium level once food supplies/cash reserves diminish or when social protection programmes cease operations. The WMS found that about 5 percent of households relied principally on obtaining piecework for their main food source (NSO, 2008) and these wage rate improvements would thus significantly reduce poverty in at least 144,000 households. These impacts are potentially seasonal and are directly correlated with the scale of the productivity gain arising from the subsidy programme. This trend needs to be considered together with the growth in real remuneration of the formal sector workforce from the late 1990s to 2006 (Durevall and Mussa, 2010).24 Rural wages (including casual wages) nevertheless remain very low, with the statutory minimum at US$0.75 per day, and many casual workers are paid less than this amount, either fully or partially in-kind (usually with food). The expansion in smallholder maize, burley tobacco, roots, tubers and legumes has absorbed available labour resources during the main cropping and post-harvest periods. Using data from research on crop labour requirements (person days), along with the MoAFS crop estimate statistics, we were able to estimate smallholder labour requirements for these crops. The estimation assumes an average monthly labour input of 26 days. As yield varies in the case of maize and tobacco, correlating with the scale of inputs, we consider high and low labour-input scenarios. The results are shown in Table 2.8. The number of persons required to cultivate the current area of land under maize on a daily basis for four months is 2,152,958 at a low input level and 2,870,611 at a high input level. For tobacco, the number of persons required for nine months at a low input level is 395,166 and 497,935 for a high input level. During the summer rainfall season, when smallholders are

Social protection, growth and employment

Table 2.8: Maize and Tobacco Labour Input Fertilizer

Labour Requirement Persondays/ hectare

Total hectares 2009/10

Total persondays

Length of season/ month

Seasonal requirement/ month

Seasonal requirement/ day

Maize high

176

1,696,270

298,543,520

74,635,880

2,870,611

Maize low

132

1,696,270

223,907,640

55,976,910

2,152,958

Tobacco high

751

155,149

116,516,899

12,946,322

497,935

Tobacco low

596

155,149

92,468,804

10,274,312

395,166

Cassava

171

195,828

33,486,588

2,232,439

85,863

Total low

349,863,032

68,483,661

2,633,987

Total high

448,547,007

89,814,641

3,454,409

Source: Person-days/hectare: Maize and Tobacco (Takane, 2008:193); Tobacco (Koester, et al., 2004:64); Maize and Cassava (Barratt, et al., 2006:118); Crop hectare: MoAFS Final Crop Estimates.

producing all three crops, the total labour requirement necessitates the full-time involvement of between 2.6 million (low input) and 3.4 million (high input) Malawians of working age in agricultural endeavours. These figures need to be seen in the context of the total potential adult (18–65 years) rural labour force, which currently amounts to about 5.5 million persons (both genders). The results show that between 47 percent (low input) 62 percent (high input) of this total potential labour force needs to be permanently engaged in agricultural activities for at least four months to achieve the current level of production in maize alone. Smallholders need to balance this demand on their labour against the labour requirements for other crops (cash crops [cotton/coffee/tea/sugar/groundnuts], grains [rice/millet/sorghum], legumes, fruit and vegetables), livestock production, natural resource harvesting, and off-farm income activities. The WMS 2008 survey findings which report that rural unemployment was less than 1 percent is consistent with the proposition that rural labour force has achieved near maximum seasonal employment during the agricultural season of maize/tobacco production. The cumulative evidence suggests that the rural labour market faces an emerging scarcity in supply, certainly for the period from October to May. The impact that other forms of social protection have had on either the price or supply of agricultural labour in Malawi—given current levels of on-farm demand—are little known. PWP intentionally influence labour markets, either through providing employment to the unemployed, through raising the work rate to enable the beneficiaries to achieve acceptable consumption levels (thus attaining household food security) or through providing assets. Although most PWP are scheduled for the off-season when the demand for agricultural labour is low,25 this has not always taken place and some programmes have competed for labour. But their scale—outside of food crises situations—has been limited and the disruption generally short in duration. There is evidence that FFW programmes have reduced household expenditure on food. In times of widespread food shortage, such as during 2001/02, FFW interventions helped to stabilize maize

Malawi

market prices (reducing the costs of maize), thus minimizing the length for which food insecure households had to engage in piece-work. While food transfers have surely had an inflationary influence on wage rates, it is reasonable to assume that the impact has been short in duration and relatively isolated. Conditional cash transfer programmes have had a direct influence on rural labour markets and could have a long reaching impact if they are sustained as proposed. In the experience of the GoM/UNICEF Social Cash Transfer Scheme (SCTC), the transfer of US$14 per month to poor households has lessoned their reliance on piece-work. However, instead of entering the labour market themselves, there is evidence that benefiting households have begun to employ piece workers in on-farm activities (Miller, 2009). Almost half the beneficiaries hired in labour. Some of this employment arises from a reciprocal gesture, whereby the beneficiaries aim to redistribute a portion of their gains to other, non-benefiting, households and thus preserve good social relations within the community. Other forms of reciprocal gestures included the intra-household distribution of gifts of food and loans. This conforms to cultural practice. As one researcher has noted, piece work employment is regarded (by both buyers and seller) as part economic contract and part moral obligation, whereby wealthier families (i.e. non-poor) are expected to provide employment to poor households (Takane, 2009: 193). In the case of the SCTC, research by Miller found that the provision of employment was the most frequently reported means through which beneficiaries redistribute benefits to non-beneficiaries. The scale of this redistribution is small in value terms (less than 5 percent of the original sum transferred), but nevertheless accounts for about half of the work opportunities that non-beneficiaries were able to obtain in this case on a monthly basis (Miller, 2009: 22).

Conclusion The structural characteristics of poverty in Malawi dictate that agriculture and investment in small-scale farmers in particular provide the main pathway for alleviating poverty. The population is overwhelming poor and entrenched in a rural economy that is based on small farms producing subsistence and cash crops. Rural politics is dominated by traditional leadership and democratic institutions are still weak. The population is fast growing and young, almost half under the age of 15. There is little formal industry, service sectors or mining to provide alternative employment. In agriculture, Malawi holds regional comparative advantages in agro-ecological conditions, including high rainfall, while the rural labour force is skilled in (traditional) crop husbandry and willing to work at relatively low wages. The Malawi case provides evidence of the reduction of poverty through the provision of maize technology transfers, initially as a result of the fee distribution of small start-up packages and later through subsidies. There is an emerging consensus that input subsidies, as a developmental tool, can â&#x20AC;&#x2DC;raise land and labour productivity and improve food security for large numbers of poor [Malawian] householdsâ&#x20AC;? (Dorward, Chirwa et al., 2010: 34). The study supports the argument that increases in smallholder maize production are the fastest way to improve food security and to reduce poverty. The scale of this achievement is notable, especially from 2005.

Social protection, growth and employment

The growth in national maize production in the period 2007 to 2010 significantly reduced the need for food aid and safety net measures to mitigate hunger. National surveillance surveys report that the number of persons living below the poverty line fell by 12 percent between 2004 and 2008 and currently (2010) stands at 40 percent. There is quantifiable evidence that the welfare benefit has been evenly distributed across all strata of poor people. Whereas the non-poor gained from increased crop sales and increased home production, poor people also increased subsistence production and gained from higher wages for casual labour. Over this brief time, poor and ultra-poor people have managed to accumulate assets, most notably livestock (goats, pigs and cattle) and also to set aside savings. On the basis of the number of households recorded in the 2008 census, the national average number of Tropical Livestock Units has risen from 0.53, as recorded in 2004, to 0.57 (a considerable achievement, given the rapid levels of population growth), with considerable expansion in the goat population.26 The increase in small livestock is important because women are culturally entitled to possess goats and pigs as well as poultry (which they do, as national surveys confirm) and the growth trend implies that they own more livestock. The main areas of benefit among the strata of ultra-poor, poor and non-poor people in this case are summarized in Table 2.9. The role of other forms of social protection, including PWP and direct transfers, in the reduction of poverty is debatable, though PWP are certainly a logical complement to crop enhancement because of complementary seasonality. But, unlike the case of productivity enhancement measures, their impact has been geographically limited. Their main beneficiaries include cohorts of poor and ultra-poor people, who have nonetheless gained directly and indirectly from input transfers and subsidies. Given the narrower objectives of some of these alternative social protection measures, it is fair to conclude that their main impact has been to smooth consumption and enable households to more easily overcome minor shocks through improved income. Their role in the empowerment of marginalized (ultra-poor) smallholders and of women, in particular, deserves acknowledgement, although it remains unclear whether this has contributed towards further on-farm investment. The case study provides important lessons on the role of agricultural productivity enhancement social protection in poverty alleviation: 1. Input transfers and subsidies were most effective when targeted broadly â&#x20AC;&#x201D; in other words, to all small-farm households â&#x20AC;&#x201D; so as to benefit the ultra-poor as well as poor and non-poor. The scale of the interventions determined their outcome. To Malawian smallholders themselves, universal targeting is considered justifiable, as all farms are small and there is no considerable wealth gap between ultra-poor and non-poor households. In terms of material possessions, for example, all rural households possess few assets (only 3 percent of households have oxcarts, almost none have motor vehicles or tractors). There has been much criticism that the targeting of AISP (and other social protection measures) has been wide of the mark, benefiting the non-poor and excluding some of the most deserving households. It is also true that there has been considerable leakage. These concerns, while legitimate, overshadow the long-term, accumulated impact that social protection measures have had over the course of the decade.

Malawi

Table 2.9: Distribution of Benefit from Productivity Enhancement Measures Type of Benefit

Social Protection Measure Ultra-Poor

Poor

Non-Poor

Land scarce, with labour

Land and labour scarce

Landless

Food

Increased maize output has reduced the scale of dependence on markets.

Marginal or no improvement in food position, though greater access to marketable supplies due to increased national food availability.

Still depend on markets (though food availability has improved) and thus have remained seasonally food-insecure when prices spike.

Increased maize output and comparatively higher yields, though still dependent on markets for a portion of their household requirements.

Increased maize output, household requirements fulfilled. Main productivity enhancement achieved in non-food crops (tobacco).

Assets

Marginal improvement, livestock accumulation

Marginal or no improvement

Livestock accumulation

Livestock accumulation, savings, and reinvestment in fertilizers

Income

Income position improved through reduced expenditure on food. Selling fertilizer would have provided additional income, but hindered their ability to invest in on-farm food production.

Wealth transfers â&#x20AC;&#x201D; including selling fertilizer and cash â&#x20AC;&#x201D; have improved household income. If able to employ labour to support on-farm agriculture, income position would have further improved. Wages remain low.

Labour demand and higher wages from piecework may have improved their position. Low labour demand in off-season remains a challenge. Some household income gain from migration and remittances. Wages remain low.

Increased household income from crop sales (estimated at between 6% and 8%). However, higher wages would have reduced their capacity to employ labour.

Increased household income from commodity crops/maize, which has improved their ability to hire in labour.

Still highly vulnerable to shocks; dependent on off-farm income, asset transfers and unable to enhance on-farm productivity.

Still highly vulnerable to shocks; dependent on scale of smallholder investment in agriculture and thus on labour demand and still reliant on asset transfers and PWP.

Household resilience strengthened through crop diversification, though farming system dependency on maize (and inputs) reinforced.

Household resilience strengthened through good returns from investment in maize/tobacco. But farming systems remain undiversified, whilst dependency on inputs enhanced.

Household Still highly vulnerable to resilience shocks, though crop diversification may have reduced food insecurity risks.

Social protection, growth and employment In a strongly traditional society, non-formal mechanisms of wealth redistribution (or reciprocal gestures) evidently fulfil a corrective action to mistargeting, ensuring that, amongst smallholders, the benefit of social protection is widely distributed (though not necessarily equally), both directly (through sharing assets/inputs) and indirectly (through providing employment opportunities). These reciprocal gestures are rooted in cultural practices, ties of clan and responsibility of the tribe (see Chinsinga, 2005). 2. Enhancing on-farm productivity from providing free or heavily subsidized inputs (seed and fertilizer) improved the comparative profitability of maize and tobacco and thus stimulated surplus production, which resulted in greater household income either directly from sales or indirectly through lessoning reliance on markets to acquire the missing household food entitlement. Furthermore, the technology transfer raised yields, thus enabling farmers to release land and labour that could then be used for other crops including cash crops and traditional food staples such as cassava, sweet potatoes and beans. 3. Input transfers/subsidies and other social protection measures have empowered women directly through enhancing their income or through capacity building and indirectly through the opportunities to expand production of crops in which their entitlement is secure. The national output of these crops exceeds the upward trends in maize and tobacco, proving that women have been able to devote more land and labour to crops that directly benefit the household (food) and themselves (income). 4. The resulting growth in agricultural production (in the targeted crops and in the indirectly affected traditional crops) fully absorbed the rural labour supply (throughout the main cropping season), creating labour shortages, which the market sought to correct through substantially raising piece-work rates. The readjustment in wages equates to a transfer of wealth from non-poor (those hiring in labour) to poor and ultra-poor people (those hiring out labour). Research has shown that ultra-poor (and even poor) people benefit as much or even more from piece-work opportunities and off-farm incomes (through a higher rate of return) than from investing additional labour in on-farm agricultural activities, since the latter provides a lower rate of return. 5. The various programmes have collectively advanced smallholder knowledge of the value of green revolution technologies through demonstrating the potential of higheryielding seed and nitrogenous fertilizer. In the case of maize, this has resulted in more widespread adoption OPV seeds and displacement of low-yielding traditional varieties. In this respect, further technical support along with inputs can realistically enable smallholders to double their current yields. It is has been demonstrated that smallholders can realistically aim to achieve yields of 4500 kg/hectare. However, research and development must be guided by a clear agricultural development strategy which should aim to achieve optimal use of the countryâ&#x20AC;&#x2122;s land and water resources, avoiding past errors in this respect such as promoting maize in unsuitable agro-ecological areas, and ensure relative price stability in strategically important food crops.

Malawi

The smallholder response to input transfers and especially to the AISP confirms the theory that small farms (as opposed to the estate sector/larger smallholders) are relatively more productive and that they are more socially efficient in using available labour resources (for the importance of this finding in the context of land reform debates, see Lipton, 2009; Berry, 2011). Despite concerns that channelling subsidized maize technology inputs to smallholders would reinforce a “vicious circle of the low productivity maize production trap” (Dorward et al., 2008), evidence from detailed field studies suggests otherwise.27 An investigation of the impact of AISP on household productivity found that the ‘relatively land-poor’ had maize yields that were 360 kg/hectare to 380 kg/hectare higher than the relatively land rich (i.e. non-poor) (Holden and Lunduka, 2010a). Moreover, the study found that relatively ‘labour-poor’ households had average maize yields about 360 kg/hectare lower than ‘labour-rich’ households, a finding that confirms our proposition that input subsidies have increased labour scarcity. The importance of raising small-farm productivity should not be understated. There is evidence from other countries that this is an effective growth-producer, having multiplier impacts on non-agricultural activities (see Berry, 2008: 5). Questions of sustainability have bedevilled the Malawian experience in input transfers and input subsidies. The concern that these programmes impose a high cost burden on the fiscal position is well founded. The AISP case shows that overall costs have escalated annually, while gains in programme efficiency through improvements in procurement and distribution (including giving the private sector a larger role) are unable to offset the spiralling trend. Further efficiency could be achieved, though this would have limited impact on the overall costs, as the predominant cost item is fertilizer. As a consequence of rising input prices, the size of the transfer and scope has had to be reduced; the programme now focuses on maize alone. While subsidization on a large scale cannot be undertaken without sustained donor financial support or the virtual erosion of the agricultural budget, our analysis confirms that it is a cost-effective measure compared to other forms of social protection. In budget terms, the best bet is for a universal subsidy that is larger than the original Starter Pack (15 kg of nitrogenous fertilizer and seed), but, on cost considerations alone, less than AISP (50 kg of nitrogenous fertilizer, 50 kg of base dressing and seed). It is important that the size of the transfer achieve more than a ‘consumption-smoothing’ impact, for, unless productivity enhancement releases land and labour, it cannot provide a pathway out of poverty. Whether smallholders have the means (and will) to purchase the same quantities of inputs at market rates will determine the sustainability of current output. Available research indicates that only 20 percent of rural households have the ability to procure a seed and fertilizer package at the full market price, while about half of households could afford a 50-percent subsidized package (Holden and Lunduka, 2010b: 33). As cost concerns come to determine the future shape of input subsidies, further reduction in fertilizer use will affect national production and household income and could reverse recent gains. Input transfers and free inputs are not a panacea. Agricultural development in Malawi requires far more investment. There is need, for example, for greater investment in public research facilities and extension services (both of which have largely been dismantled) to promote a productivity increase in traditional food crops through appropriate technology adoption, focusing in particular on disseminating low-cost seeds to offset dependence on patented

Social protection, growth and employment

varieties. Actions to further empower women and to strengthen their entitlements in traditional crops and livestock should be intensified, alongside incentives to encourage family planning so as to reduce the high fertility rate. Savings should be promoted, in order to provide farmers with an option outside of investment in small livestock (whose growing numbers will put pressure on environmental resources); expanding access to banking should be the first step in this process. The major agricultural implement in most households remains the hoe, with land preparation and weeding done by hand. Most smallholders make no use of irrigation, apart from winter cultivation of wetlands. There is also negligible use of mechanization or animal draught. So long as smallholder farming systems remain dependent on labour, which has hitherto been relatively cheap and widely available, there remains little incentive to adopt modern technologies. For agriculture to maintain its role in driving economic growth and reducing poverty, an appropriate basket of social protection measures should include: a universal inputs transfer to poor and ultra-poor people who have land and are able to work to increase household maize production; an inputs subsidy for non-poor people (reducing the costs by at least half) for investment in either maize or commodity crops; and food and cash transfers to specific cohorts of ultra-poor people—especially the landless—households unable to work and those households scare in labour. The government endorses this strategy, as expressed in its 2009 Social Support Policy. But whether it can stay the course and convince donors to stay on board is a political question, subject to domestic political pressure (from poor people, who do not want to see their gains reversed, and the non-poor, who want to retain their position as beneficiaries) and donor equivocation over policy direction and their strategic manoevering to obtain a controlling influence in programme implementation.

Notes 1 Data from hdrstats.undp.org/en/countries/country_fact_sheets/cty_fs_MWI.html . 2 There is a small manufacturing sector in Malawi that is mainly concentrated in agro-processing activities, especially in tobacco, tea and sugar. Import substitution manufacturing is in decline but includes themanufacturing of detergents, candles and basic pharmacological commodities. The clothes industry has largely collapsed, and, as Malawi is a landlocked country, access to export markets is severely constrained. 3 The WMS defines labour force participation as the proportion of the population 15 years and above who are employed and unemployed (looking for and ready for work). 4 The president’s decision to exhaust the country’s limited foreign exchange reserves to procure a presidential jet is one example of these failings. See en.wikipedia.org/wiki/Bingu_wa_Mutharika . 5 The International Finance Corporation’s Doing Business 2011 report notes that Malawi has made important steps towards improving the business environment through easing the process of property transfers and improving opportunities for the enforcement of contracts in lower courts (2010: 139), but barriers in some sectors (such as construction permits) are onerous by global comparison. 6 This benchmark is considerably lower than the UNDP poverty line of US$1.25 per capita per day.

Malawi

7 It should be noted that the notion that the poor are unable to make rationale economic decisions between labour market participation (and other forms of off-farm income) and on-farm production, as part of this entrapment, has been dispelled (see Orr et al., 2009). 8 Migrants provide an import source of income to their families, transferring funds within the country and from abroad. The scale of remittances through formal financial institutions from outside the country is relatively small, estimated at about US$1 million per year (Durevall and Mussa, 2010: 38). The Finscope study found that half of the Malawians who held bank accounts (about 20 percent of the adult population are banked) needed such accounts to receive remittances. Around 15 percent of Malawians reported that they either receive or send money within the country each month (Finscope, 2009). 9 This subsection draws upon the writings of Mkandawire (1997), Harrigan (2003), Masanjala (2006) and Peters (2006). 10 The 2001/2002 crisis was not simply a case of diminishing resilience. Maize output in 2001/2002 was low, but did not amount to a national crop failure and was consistent with the previous 10-year average. This has led analysts to point to the institutional and market failures that impacted on household access to maize. See Devereux (2002), Devereux and Tiba (2007) and Charman (2004). 11 In the reports from the WFP, this means that food aid was provided to 23 million clients, with some individuals receiving food aid regularly for several months. 12 There is disagreement on the scale of the impact of cash transfers on maize markets; some scholars argue that these programmes have inflated food prices, thus worsening vulnerability. 13 In reality, PWP undertaken by INGOs tended to offer much higher wages. 14 Data from 2008 Report on the global AIDS epidemic, UNAIDS/WHO, July 2008. Annex 1: HIV and AIDS estimates and data, 2007 and 2001 and WHO Report on Tuberculosis 2008. 15 Social protection and TIP would justify the continued existence of ADMARC at a time that the IMF/ World Bank were pushing for complete privatization of input markets. 16 Several donors, including the EU, simultaneously supported INGOs to advance low-cost input solutions for poor smallholders. 17 In terms of the existing institutional architecture, the leading donors (World Bank, DFID and the EU) were pushing for ‘rationalization’, reducing state services to their ‘core function’ and allowing the private sector to replace the state as the main channel for technology transfer. 18 At the same time, numerous INGOs were advancing a wholly different approach, focusing on raising on-farm productivity, placing the emphasis not on inorganic fertilizers and hybrid seeds, but on low-cost, adaptable and renewable technologies. For one articulation of this counter-argument, and its current relevance to the situation, see Grain (2010). 19 For example, ‘Malawi sexed up the maize estimates—US study: maize supplies may be depleted before harvest time; point at inaccurate crop estimates as the main cause’ (www.tinyurl.com/malawi-maize). 20 One of the most ‘successful’ initiatives in promoting OPV seed was the GoM/EU Promotion of Soil Conservation and Rural Production Project (PROSCARP), a national intervention through the agricultural extension service.

Social protection, growth and employment

21 Ironically, the MVPA cautions against this strategy, arguing that “low-value traditional crops” are likely to “trap” smallholders at “a bare subsistence level’” (GoM, 2006: 71). 22 The analysis does not adequate reflect the influence all significant variables, including displacement costs (probably minor), mistargeting and leakages (less significant, especially under universal targeting), and the incremental grain production gain from the use of hybrid/OPV seed and inorganic fertilizer. 23 This thinking is applicable to food aid, not to commercial imports. 24 Within firms in the agricultural sector, for example, wages rose by 150 percent over this period, partially in response to currently devaluation. 25 Rural people use the off-season to catch up on livelihoods activities and asset building (including repairing and building houses), work activities that could not be undertaken during the agricultural season due to the intensity of the labour demands. 26 Estimate calculated using the following conversions: cattle = 0.7, small ruminants = 0.10, pigs = 0.2, and poultry = 0.01. livestock data from FAOSTATS. 27 While the authors recognize that free inputs/subsidies have contributed towards lowering and stabilizing maize prices and raising productivity, they are in favour of “encouraging less people to grow maize, but to grow it more productively” (Dorward et al, 2008: 19).Yet the experience from 2005 seems to indicate that some of the very smallholder that, in these terms, ought not to grow maize, have in fact made significant productivity gains. Moreover, the strategic emphasis on maize (which has undoubtedly intensified since 1994) has not detracted from smallholder (and female) efforts towards crop diversification. Public sector extension support for regional specialization in crops other than maize must, in any case, support these emerging trends.

Malawi

References Alwang, J. and P.B. Siegel, ‘Labor shortages on small landholdings in Malawi: Implications for policy reforms’, World Development, 1999, 27(8): 1461–1475. Barratt, N., D. Chitundu et al., ‘Cassava as drought insurance: food security implications of cassava trials in Central Zambia’, Agrekon, 2006, 45(1). Berry, A., ‘What type of global governance would best lower world poverty and inequality?’, 2008. Berry, A., ‘Review of Michael Lipton Land reform in developing countries: property rights and property wrongs’, Development and Change, 2011, Vol. 42, No. 1. Binswanger, H.P. and J. McIntire, ‘Behavioral and material determinants of production relations in land abundant tropical agriculture’, Economic Development and Cultural Change, 1987, 36(1): 73–99. Blackie, M. and C.K. Mann, ‘The origin and concept of the Starter Pack’, Starter packs: a strategy to fight hunger in developing countries? Lessons from the Malawi experience 1998–2003, 2005: 15–27. Charman, A.J.E., ‘Agricultural development and food security in sub-Saharan Africa (SSA): building a case for more support. The case of Malawi’, Policy Assistance Series Working Paper, No. 4, 2004. Charman, A.J.E., ‘Empowering Women Through Livelihoods Orientated Agricultural Service Provision’, UNU-WIDER Research Paper, No. 2008/01, 2008. Chilimampunga, C., P. Hopcraft et al., Consultancy Report. Managing the emerging large farm crisis and options for land resettlement, 1998. Chinsinga, B., ‘Poverty and food security in Malawi: Some policy reflections on the context of crumbling traditional support systems’, Canadian Journal of Development StudiesRevue Canadienne D’Études Du Developpement, 2004, 25(2): 321–341. Chinsinga, B., ‘The clash of voices: Community-based targeting of safety-net interventions in Malawi’, Social Policy & Administration, 2005, 39(3): 284–301. Chinsinga, B., ‘The social protection policy in Malawi: processes, politics and challenges’, Future Agricultures Policy Brief, No. 24, Institute of Development Studies, Brighton, 2007. Chinsinga, B., ‘The Malawi Fertiliser Subsidy Programme: politics and pragmatism’, Future Agricultures Policy Brief, No. 022, Institute of Development Studies, Brighton, 2008.

Social protection, growth and employment

Chirwa, E., P. Mvula et al., The evaluation of the Improving Livelihoods Through Public Works Programme, Malawi Social Action Fund, 2004. Davies, S. and J. Davey, ‘A regional multiplier approach to estimating the impact of cash transfers on the market: The case of cash transfers in rural Malawi’, Development Policy Review, 2008, 26(1): 91–111. Devereux, S., ‘The Malawi famine 2002’, IDS Bulletin, Institute of Development Studies, Brighton, 2002, 33(4): 70. Devereux, S., B. Baulch et al., ‘Vulnerability and social protection in Malawi’, Discussion Paper, No. 387, Institute of Development Studies, Brighton, 2006. Devereux, S. and Z. Tiba, ‘Malawi’s first famine, 2001–2002’, Routledge, London, 2007. Dorward, A., E. Chirwa et al., ‘Towards “smart” subsidies in agriculture? Lessons from recent experience in Malawi’, Natural Resource Perspectives, 2008 (116). Dorward, A., E. Chirwa et al., ‘The Malawi Agriculture Inputs Subsidy Programme, 2005/6 to 2008/9’, International Journal of Agricultural Sustainability, 2010, 9(1): 232–247. Dorward, A., B. Guenther et al., Linking social protection and support to small farmer development. Malawi Case Study, FAO, Rome, 2008. Dorward, A. and J. Kydd, ‘The Malawi 2002 food crisis: the rural development challenge’, Journal of Modern African Studies, 2004, 42(3): 343–361. Durevall, D. and R. Musaa, Employment diagnostic analysis on Malawi, Government of Malawi, Ministry of Labour, 2010. FEWSNET, Malawi Food Security Update, November 2009. FinScope, FinScope Malawi 2008, Finscope, 2009. Food and Agriculture Organization of the United Nations (FAO), Special Report. FAO/WFP Crop and Food Supply Assessment Mission to Malawi, FAO, Rome, 2005. Government of Malawi, Malawi Poverty Reduction Strategy Paper, Government of Malawi, Lilongwe, 2002. Government of Malawi, Malawi Growth and Development Strategy: From Poverty to Prosperity 2006–2011, Government of Malawi, Lilongwe, 2006. Government of Malawi, Social Support Policy, Government of Malawi, Lilongwe, 2009. Government of Malawi, A medium term plan for the Farm Inputs Subsidy Programme, 2011–2016. Third Version, Ministry of Agriculture and Food Security, Government of Malawi, Lilongwe, 2010.

Malawi

Government of Malawi and World Bank, Malawi poverty and vulnerability assessment: investing in our future, Government of Malawi and World Bank, Lilongwe, 2006. Harrigan, J., ‘U-turns and full circles: Two decades of agricultural reform in Malawi 1981–2000’, World Development, 2003, 31(5): 847–863. Harrigan, J., ‘Food insecurity, poverty and the Malawian Starter Pack: Fresh start or false start?’, Food Policy, 2008, 33(3): 237–249. Hodge, J. and A.J.E. Charman, ‘Analysis of the Potential Impact of the Current WTO Agricultural Negotiations on Government Strategies in the SADC Region’, Food security: indicators, measurement, and the impact of trade openness, 2008: 239. Holden, S. and R. Lunduka, Impacts of the fertiliser subsidy programme in Malawi: targeting, household perceptions and preferences, Norwegian University of Life Sciences, As, 2010a. Holden, S. and R. Lunduka, Too poor to be efficient? Impacts of targeted fertiliser subsidy programme in Malawi on farm plot level input use, crop choice and land productivity, Department of Economics and Resource Management, Norwegian University of Life Sciences, As, 2010b. International Monetary Fund (IMF), Malawi: Debt Relief at the Heavily Indebted Poor Countries Initiative, IMF, Washington, DC, 2006. International Organization for Migration (IOM), Briefing note on HIV and labour migration in Malawi, IOM, Geneva, 2006. Kerr, R. B., ‘Food security in northern Malawi: Gender, kinship relations and entitlements in historical context’, Journal of Southern African Studies, 2005, 31(1): 53–74. Kerr, R. B., ‘Informal labor and social relations in northern Malawi: The theoretical challenges and implications of ganyu labor for food security’, Rural Sociology, 2005, 70(2): 167–187. Kerr, R. B., ‘Groundnuts as “economic crop” or “wife of the home” in Northern Nyasaland’, Journal of Historical Geography, 2010, 36(1): 79–89. Koester, U., G. Olney et al., Status and prospects of Malawi’s tobacco industry: a value chain analysis, The Emergency Drought Recovery Project, Ministry of Agriculture, 2004. Levy, S., Financing and macro-economic impact: how does Starter Pack compare?, CABI Publishing, Wallingford, UK, 2005. Levy, S., C. Barahona et al., ‘Food security, social protection, growth and poverty reduction synergies: the starter pack programme in Malawi’, ODI Natural Resource Perspectives, 2004, (95): 6.

Social protection, growth and employment

Lipton, M., Land Reform in Developing Countries: Property Rights and Property Wrongs, Routledge, London, 2009. Malawi Vulnerability Assessment Committee (MVAC), Food Security Monitoring Report, MVAC, Lilongwe, 2005. Malawi Vulnerability Assessment Committee (MVAC), Malawi’s 2007 Bumper Harvest: Is everyone food secure?, MVAC, Lilongwe, 2007. Malawi Vulnerability Assessment Committee (MVAC), ‘Fifth food surplus year in a row ... but dry spells mar access’, The Malawi Vulnerability Assessment Committee Bulletin, 2010, 6(1). Masanjala, W.H., ‘Cash crop liberalization and poverty alleviation in Africa: evidence from Malawi’, Agricultural Economics, 2006, 35(2): 231–240. Miller, C., Economic impact report of the Mchinji social cash transfer pilot, Center for International Health and Development, Boston University School of Public Health, Boston, MA, and The Centre for Social Research, University of Malawi, Lilongwe, 2009. Miller, C.M., M. Tsoka et al., ‘Targeting Cash to Malawi’s Ultra-Poor: A Mixed Methods Evaluation’, Development Policy Review, 2010, 28(4): 481–502. Ministry of Agriculture and Food Security, Final Round Crop Estimates, Ministry of Agriculture and Food Security, Lilongwe, 2010. Ministry of Agriculture and Livestock Development, The Agricultural and Livestock Development Strategy and Action Plan, Ministry of Agriculture and Livestock Development, Lilongwe, 1995. Mkandawire, T., ‘Malawi agriculture, poverty and employment’, unpublished document, 1997. Msachi, R., L. Dakishoni et al., ‘Soils, food and healthy communities: working towards food sovereignty in Malawi’, Journal of Peasant Studies, 2009, 36(3): 700–706. National Statistical Office (NSO), Integrated household survey 2004–2005. Volume I, NSO, Lilongwe, 2005. National Statistical Office (NSO), Main report census 2008, NSO, Lilongwe, 2008a. National Statistical Office (NSO), Welfare Monitoring Survey, NSO, Lilongwe, 2008b. Ntata, P.R.T., ‘Bridging the hunger gap with cash transfers: experiences from Malawi’, Development in Practice, 2010, 20(3): 422–427. Orr, A., B. Mwale et al., ‘Exploring Seasonal Poverty Traps: The “Six-Week Window” in Southern Malawi’, Journal of Development Studies, 2009, 45(2): 227–255.

Malawi

Peters, P.E., ‘Rural income and poverty in a time of radical change in Malawi’, Journal of Development Studies, 2006, 42(2): 322–345. Regional Hunger and Vulnerability Programme (RHVP), Dowa Emergency Cash Transfer, Malawi. REBA case study brief, RHVP, Johannesburg, 2007a. Regional Hunger and Vulnerability Programme (RHVP), Input Subsidy Programme, Malawi, RHVP, Johannesburg, 2007b. Regional Hunger and Vulnerability Programme (RHVP), Public works programmes, Malawi. REBA case study brief, RHVP, Johannesburg, 2007c. Reserve Bank of Malawi, Financial and Economic Review, 41, Reserve Bank of Malawi, Liongwe, 2009. Reserve Bank of Malawi, Monthly economic review, June 2010. Rockefeller Foundation, Malawi: Soil Fertility Issues and Options. A discussion paper, Rockefeller Foundation, New York, 1998. School of Oriental and African Studies, Wadonda Consult et al., Evaluation of the 2006/7 Agricultural Input Subsidy Programme, Malawi, School of Oriental and African Studies, London, 2008. Smith, W.J., ‘Spending on Safety Nets for the Poor: How Much, For How Many? The Case of Malawi’, Africa Region Working Paper Series, No. 11, World Bank, Washington, DC, 2001. Statistics South Africa, Tourism and Migration, Statistics South Africa, Pretoria, 2010. Takane, T., ‘Labour use in smallholder agriculture in Malawi: six village case studies’, African Studies Monograph, 2008, 29(4): 183–200. Weyl, U., Labour migration and underdevelopment in Malawi, late 19th century to mid 1970s, Belmont Press, 1991. Whiteside, M., ‘Ganyu labour in Malawi and its implications for livelihood security interventions—an analysis of recent literature and implications for poverty alleviation’, AgREN Network Paper, 2000, (99): iv + 10. World Bank, Malawi. Social protection status report, Report No. 40027-MW, World Bank, Washington, DC, 2007. World Food Programme, 2005/2006 Biennial Report. Malawi, World Food Programme, Rome, 2006.

3 Addressing Transitory and Chronic Poverty: Social Policies in Peru 1990–2010 Janina Leon Introduction In recent decades Latin American governments, whether enjoying rapid growth or suffering weak macroeconomic performance, have accompanied their more narrowly economic programmes with ‘social policies’ mainly designed to deal with poverty. Peru implemented a number of such policies in the early 1990s under a crisis scenario and has continued to apply them or variants of them during the more recent period of high growth. This study reviews selected social policies and programmes in the context of Peruvian economic development, stressing the role played by the broadly defined social policies implemented mainly since 1990. Peru has historically been one of Latin America’s lower-middle-income countries, whose main links to the international economy have traditionally been via mineral or other natural resource exports. Both its relatively low income and dependency on mineral exports have contributed to a high share of the labour force (given per capita income) being occupied in informal activities and in services.1 Various disincentives to agriculture (including the focus on mineral exports) promoted an earlier than normal process of urbanization. The volatility of mineral exports ensured that the domestic economy would also suffer from severe fluctuations. These in turn have contributed to wide swings in the political sphere and in economic policy, including a number of major swings in policy orientation over the last half-century, moving from more or less populist to more or less conventional (Dornbusch and Edwards, 1991). Although the country

100

Social protection, growth and employment

has had the benefit of a number of excellent economists and technocrats, institutional quality overall has tended to be weak until fairly recently (Thorp, 1991). Left-of-centre policies, such as the worker management system, never had the potential to involve more than a quarter of the population, since the great majority were employed in small agriculture or the urban informal sector (Webb, 1975). This complicated scenario has provided the backdrop for the evolution of economic and social policy in Peru. Until the 1980s, social policy was traditional and managed and funded by the state, with the resulting subsidies tending to reach only those who were well informed and organized. By 1990 the operative definition of social policy had been narrowed to focus on specific goals and on strategies for specific poor populations.

Social policy and social protection in public policy2 Social policies in general and social protection polices more specifically reflect each government’s preferences for how to deal with societal needs in areas such as health and education and to respond to the specific needs of poor people; usually they are not connected closely, if at all, to the economic policies implemented. In some cases, government is the sole provider and manager of the social policies; at the other extreme, market mechanisms may be used with government regulation; intermediate policies involve both public and private organizations. Designing programmes and tools to effectively reach low-income populations is a challenge. In Peru in 1990 the programmes making up the social safety net suffered from many problems, including weak administrative organization, with correspondingly low cost-effectiveness and generally unimpressive results. There was a leaning toward social instruments that were self-targeting (e.g. public works hiring at below-market wages, consumption subsidies for specific poor groups such as single mothers). Monitoring and evaluation of the social safety net programmes was practically non-existent, in part because of a lack of appropriate data to undertake such analysis. Most inferences were based on extrapolation from programme targets, conditionality indicators and monitoring standards for benchmark and performance references, as the International Monetary Fund (IMF) points out (2000: 19–20).

The Peruvian experience until the late 1980s Peru’s economy has historically been primarily export-oriented, with the accompanying vulnerability and volatility, and with a small share of total employment located in these activities. Mining and other primary products were mainly produced by foreign firms isolated from the rest of the economy. In the 1960s Peru was a geographically disconnected and highly unequal society, with most of its population working in low-productivity activities—subsistence agriculture and the urban informal sector. Velasco’s 1968–1975 military government responded with an industrialization strategy, economic reforms and a push toward cooperatives and state ownership. Although this experiment coincided initially with fast growth (6 percent) and falling unemployment as a drawing

Peru

101

down of foreign exchange reserves raised capacity utilization, when that exchange ran out, inflation exploded to 100 percent, the external debt increased dramatically, and the experiment showed itself to be unsustainable. By the late 1970s another military government (Morales) confronted a high external debt, declining exports, increasing food and input imports and low foreign investment, with resulting internal economic crisis. Under the elected government of President Belaunde (1980–1985) the economy also suffered natural shocks, and the terrorist groups that subsequently devastated Peru socially and economically got their start. Short-term fiscal discipline was applied, followed by slow gross domestic product (GDP) recovery and little growth of formal-sector employment. In 1985–1987 President Garcia’s first term once again instituted a major policy shift—raising the protection of national industry, imposing controls on foreign trade, introducing multiple exchange and interest rates and price controls, and establishing a foreign debt payment ceiling. Wage increases outran internal production as local industry grew slowly, with resulting inflation and devaluation, along with slow GDP growth and increasing foreign debt. By mid-1987 two main interventions, bank nationalization and curtailment of external debt servicing, brought political and economic retaliation from abroad. By late 1989 the Peruvian economy was suffering a deep recession (GDP shrinking by 9 percent), fiscal and external deficits, worsening external isolation and hyperinflation (more than 5000 percent per year). Terrorism and drug trafficking added their negative effects to the mix. Poverty and extreme poverty grew, inequality increased (Saavedra and Diaz, 1999: 30), and provision of public services (education and health) worsened. Labour was challenged by this scenario, which led to increased internal migration to urban and jungle areas, emigration of middle-class workers and growing informality. The increase of urban poverty accelerated; according to Escobal, Saavedra and Torero (1998: 7) the urban poverty rate changed from 39 percent in the 1970s to 36 percent in 1985 and 53 percent in 1991. Among the 1960s programmes with at least some presumed ‘social’ impact were large public investments (roads in the jungle and the Andes) financed in part through international cooperation (e.g. USAID, Peace Corps, USA Alliance for Progress), as well as monetary and in-kind transfers for poor people. Most public investment went to urban areas, where most Peruvians (including the workers already there and political and social movements, including new political parties) expected the country’s modernization to occur, and where unionization was strong. As part of its policy of full coverage of all so-called ‘basic needs’ (including all types of health, education and housing services), the 1970s military government provided public goods such as schools and large hospitals as part of its medium- and long-term plans in those areas, as well as infrastructural spending on major roads. Its other main social policy instruments included a high minimum wage, support for unionization, profit-sharing with employees by private firms, rising public-sector employment, public expenditure on housing, and social security provided through formal-sector firms. Poverty and equality objectives were to be pursued not primarily through transfers but by changing property rights towards collective ownership in both urban and rural areas.

102

Social protection, growth and employment

With the crisis of the late 1980s, when GDP fell by an accumulated 24 percent (1988–1990), most of these programmes fell victim to a lack of funding. Urbanization accelerated over 1985–1990, but waged employment fell as a proportion of the total, the real wage decreased, poverty increased, especially in rural areas and smaller cities, and social tensions increased. The critical economic and social scenario of 1987–1990 not only forced an end to ‘populist’ macroeconomic policies but also led to a reduction in social spending, which changed from 3.1 percent of GDP in 1980 to 2.5 percent in 1985, 3.7 percent in 1987 and 2.9 percent in 1991 (Vasquez et al., 1999: 642). Also, price controls on goods and subsidies for public utilities were weakened. Among the modest-sized social programmes were: Presa (Programa de seguridad alimentaria) for the poorest peasants, to assure markets and income for their production; Programa de asistencia directa for young people; Comedores populares, Glass of Milk (Vaso de leche) and other local programmes; and Programa de apoyo al ingreso temporal, mainly for urban young people (Parodi, 2000: 230–238).

1990–2010: structural reforms, poverty and employment Beginning in 1990, the Peruvian economy shifted from its previous path to a very different one. Structural adjustment policies were implemented, and, though subsequently reformulated, they have not been abandoned since then (see Annex 3.1). Among these policy shifts was a major reform to eliminate or reduce the labour regulations seen as impediments to a freely competitive labour market.3 Broadly speaking, the anticipated macroeconomic outcomes have been attained, with the economic upturn followed by solid average growth of 5.3 percent (from 1992, the first year of significant growth, to 2010),4 despite marked growth fluctuations; per capita gross national income (GNI) grew by an average of 3.6 percent per year. Other challenges have been met less successfully during this period. Poverty rates, though increasing as per capita GNI stagnated over 1997–2001, did fall sharply over the last decade to reach the lowest levels on record, with the improvements proportionately much stronger in urban areas and an increasing relative gap between rural and urban poverty. Modern waged employment increased little, while the informal sector remained very large, with a high share of workers operating in low-productivity urban microenterprises (Rodriguez and Higa, 2010). By 2008, labour in microenterprises counted for around 50 percent of total workers in the country, according to Rodriguez and Higa (2010: 6). This pattern has called into question some of the expectations and predictions of the supporters of the cited labour market reform. As detailed in the following sections, the approach to social policy changed markedly in the early 1990s, after which the norm has been minor reformulations mainly directed at management/implementation issues. Before identifying the key challenges faced by the social policies of these two decades, it is important to highlight some features of the Peruvian society and economy during this period, some due to the structural features noted above and some to the country’s level of development. They include:

Peru

103

• a relatively high level of urbanization, with 75.9 percent of the population residing in urban areas, according to the Population Census of 2007. Terrorism and other shocks help to explain the rapid migration of workers to larger urban areas, as do weak employment options in rural areas and smaller cities, based on low rates of investment and limited demand for labour by large firms in the important extractive sector; • low open unemployment rates, although hidden unemployment appears to be significant; • a low proportion of workers employed by large private firms; such firms, located mainly in urban areas, tend to hire males with considerable education and training (Rodriguez and Higa, 2010: 6); • more than 50 percent of employed workers nationwide and around 50 percent in urban areas work in microenterprises, either self-employed or running firms of up to four workers; • after the 1989 hyperinflation and the reforms of the 1990s and later, a low and stable level of real wages (Diaz, 2009) and of the legal minimum wage, which was set with a view to keeping labour costs low and making firms competitive; • high poverty rates, especially in rural areas and marginal urban areas; • a high level of inequality, with deep historical roots and associated with a very unequal distribution of wealth; inequality is higher in urban than in rural areas; and • prior to 1990, a relatively costly (to the employing firms) body of labour legislation (Riveros, 1989). In this context, one policy that deserves special attention is the labour reform which began in 1991. Its main goal was to increase firms’ flexibility, consistent with the new neoliberal economic perspective of economic policy at the time. It allowed formal firms to undertake temporary hiring, reduced the restrictions on firing, allowed subcontracting of third parties to perform tasks that were not the firm’s main activity, and lowered the obligations for non-wage benefits. One main component was reform of the formal sector’s employment-related pension system, in part because of its de facto bankruptcy, the result of previous governments using its accumulated funds to cover their fiscal deficits. Under the reform, two systems were created, one private and mostly bank-owned (AFPs) and another managed by the public sector (ONP), with both based on worker contributions. Another labour-related policy was the reduction of publicsector employment through incentives and dismissals, as a way of cutting fiscal expenditures. By the 2000s these policies had contributed to fewer formal-sector workers as a proportion of total employment, as well as a fall in unionization and other forms of labour association. The combination of the drastic economic crash of 1988–1990 and these policy measures put severe pressure on employment and led to a sharp drop in labour incomes and a rapid growth of microenterprise. While the precise impacts of the 1990s labour reform remain uncertain, it is clear that it did not lead to a large and quick reduction in the size of the informal sector by increasing the flexibility of formal-sector labour contracting. As noted by Chong et al. (2008), Peru not only

104

Social protection, growth and employment

maintained one of the highest levels of informality in the region but also experienced a consistent growth in the share of informal employment not only during the crisis of the late 1980s but also during part of the growth recovery period of the 1990s. These authors (ibid.: 244) attribute this outcome in part to changes in the orientation of the economy and technological changes related to the opening of the economy. Many older workers, more typically found in the formal sector, lost their jobs as a consequence of public-sector downsizing, privatization and the contraction of formerly protected manufacturing sectors and were unable to find new formal-sector jobs. The expectation of a significant increase in formal-sector employment as a consequence of the reform seems in any case to have been unrealistic in light of the relatively narrow range of employment on the margins of formality and susceptible to being formalized by such legislative changes. In the context of a recently proposed law designed to encourage the formalization of microand small enterprises, Rodriguez and Higa (2010: 18–19) note that 70 percent of all informal enterprises have only one person, and of those with more, only a quarter contract labour, from which they conclude that at most 6 percent might be candidates to formalize under such a law.5 While the labour reform did not produce the benefits hoped for by some, it remains unclear whether it is responsible for some of the disappointing labour market outcomes in Peru. While informality did rise until about 2000, it has since then fallen somewhat and appears more or less in line with that of other Latin American countries, when per capita income is taken into account (Morales et al., 2010: 59). It is thus not impossible that the reforms have given a small and delayed impulse to formality. Income inequality may have fallen over the past decade as well. A best guess, though, is that the reforms did not have major impacts on either of these trends. Meanwhile, growth has been solid, not spectacular by international standards—with no 10-year average exceeding 5.5 percent—but far better than the historical norm for Peru.

Social policy since 1990: a new agenda and issues in recent decades In Peru the package of policies targeting poor and extremely poor people has changed considerably in the last two decades or so, with important shifts in approach, programmes, targets, tools and evaluation. As elsewhere, the generally accepted meaning of the term ‘social policy’ has itself evolved, reflecting shifting views on the role of government in providing basic public goods for poor people. The specific basket of public goods to be provided and the specific social policy instruments and their related goals, resources, tools and target population have all been linked to the (political) approach adopted by each government to deal with poverty, satisfying basic needs, and inequality. Before 1990, Peruvian governments emphasized education, health and infrastructure as part of their broad responsibilities, along with macroeconomic management; a distinct set of ‘social’ policies was designed to deliver food and basic consumption goods to poor people in urban and rural areas. Until the 1980s, social policy was seen mainly as a universal tool to deal with emergencies and poor people’s urgent needs, while education, health and infrastructure

Peru

105

were conceived as public goods contributing to long-term economic growth. By 1990, along with the macroeconomic reforms implemented to face the economic crisis and hyperinflation, a new social policy approach was being developed, with narrower priorities, targets, programmes and tools. The main immediate objective was to deal with the unavoidable short-term negative effects of the crisis itself and the macroeconomic adjustment policies designed in response to it, especially the effects on poor people through the drastic reduction of their incomes and consumption and employment options. Thus the social policy package was designed to target, within the broader poor population, the victims of these negative shocks, by smoothing the (hopefully) short-term negative effects in the expectation of positive medium- and long-term benefits from economic recovery and growth. During 2000â&#x20AC;&#x201C;2010 this new social policy orientation was maintained with minor reformulations. Although the capacity of this social policy package to generate employment for poor people has usually been assumed to be positive, there are no mechanisms that guarantee a significantly positive impact. The policies generally provide basic consumption goods and temporary income transfers, and it depends on the nature of each programme whether it has production and/or employment effects. (It is more commonly assumed that policies towards microenterprises and microfinance, especially when managed not under a social welfare but a developmental perspective, will have such benefits.) Social policies are also affected by the political preferences of each government, which may or may not include employment as a main priority. By mid-2000 more than 80 social programmes were operated by different public-sector offices, from the central government (through its various ministries) to regional and local governments. The inter-ministerial commission (ST-CIAS) was tasked with overseeing and merging the social programmes to improve their management and resource effectiveness. Several empirical studies have identified targeting problems (i.e. leakages and under-coverage) as well as corruption. After transforming and merging some old programmes, creating some new ones and streamlining management, as of late 2010 there were 26 social programmes run by the public sector.

Social policies in the 1990s â&#x20AC;&#x201D; an overview Since the early 1990s the public authorities have been active in designing and implementing social programmes in response both to the reality of widespread poverty during the economic crisis but also to various other indicators of social discomfort. The terrorism experience of the 1980s highlighted the dangers of such social and political discomfort. In the 1990s the new social policy approach redefined its components to include public goods such as education and health. Moreover, instead of emphasizing universal coverage, it focused on transferring basic goods and services to specific target groups of the poor population (see Annex 3.2). Between 1990 and 1993, the new social policy approach was implemented under the management, public resource funding and supervision of just one authority, the Presidency

106

Social protection, growth and employment

of Ministers’ Council (PCM). The main idea was to involve all ministries in diagnosing and designing specific social programmes, while the PCM assumed overall responsibility for them; one notable step was the creation of the Cooperative Fund for Social Development (Fondo de cooperacion para el desarrollo social—Foncodes). In 1993–1995, guided by the Targeting Programme for Basic Social Expenditure, authority was decentralized from the PCM to various ministries, with the expectation of facilitating the planned increase in publicly funded social expenses and contributing to the development of organizational and management capabilities to improve implementation of new programmes. During 1995–2000, inspired by the agreements of the World Summit for Social Development (Copenhagen), President Fujimori redefined the social policy challenge as reducing extreme poverty from 20 to 10 percent by 2000, reassessing the effectiveness of basic social expenditure and achieving better targeting to reach the population unable to buy an adequate amount of food. An emphasis on women was highlighted by the creation of a special ministry, the Ministerio de promocion de la mujer y desarrollo humano. Key statistical tools (national surveys about life conditions, a national population census including data on height and weight) were developed and used to map the population in extreme poverty across more than 400 districts and to better focus the increasing social expenditure. Parodi (2000: 175–182) notes that during the 1990s social expenditures remained low as a share of total public spending and decreased on a per capita basis and as a share of GDP such that during the crisis of the early to mid-1990s the social expenditure could not compensate for the income losses of most of the population. Growth of enrolment in public schools was slow, allocation of the limited resources to public education was inefficient (favouring higher levels of education and urban areas), and the continuing low real wages of public school teachers and the poor physical infrastructure of schools contributed to a low quality of education and a still significant official adult illiteracy rate of 10 percent in 2007 (World Bank databank), with functional illiteracy somewhat higher. The public system provided most of the health services on a budget of 1 percent of GDP, and there was a heavy geographic concentration of expenses in Lima; meanwhile, private health systems ran one third of Peruvian hospitals, mainly for the benefit of a relatively rich group making up only 2 percent of the population. Other economic and institutional policy changes included the above-mentioned labour reforms implemented during 1990–1995; the one-time monetary incentives for public employees to enter voluntary retirement; the reform of the Peruvian social (public) insurance system to unify health service provision to its affiliated (mostly formal-sector) workers and to authorize private firms to operate Health Provider Enterprises (EPSs) supplying private hospital health services to workers (and their relatives) willing to pay for better health plans at market prices; creation of a private pension system (AFP) parallel to the existing public system (ONP), with individual workers free to choose either during the transition period, and with low but increasing acceptance of the private system, which as of 1999 consisted of five firms; and creation of the compensation for work service (CTS) as a system of forced savings by workers, deposited in a financial institution and available for withdrawal during later bouts of unemployment (hence

Peru

107

a personalized form of unemployment insurance). These various changes may have influenced the outcomes, although a lack of evaluation leaves it unclear how much.

Redefining social policy during the 2000s Paniagua’s short transitional government of 2000–2001 promoted institutionalized participation of civil society in designing and monitoring social policies and took the first steps to merge the multiple and isolated social programmes run by non-linked public offices. Social policy faced new challenges and experienced further changes in the decade to follow. Over 2001–2010, the two Presidents (Toledo and Garcia), though having different political leanings, pursued broadly similar social policies. Under Toledo (2001–2006), most of the previous social programmes were preserved, albeit with some changes. As of 2001, poverty and extreme poverty were still quite high6 — probably equal to their level of several decades earlier—economic growth had not been enough to bring per capita income above the level of the mid-1980s, and evidence suggested that the resources allocated to social programmes during the 1990s had mainly benefited urban and middle-income families. The large number of programmes and public institutions involved made reform difficult and slow. The main steps included: a merging of social programmes; restructuring of one ministry (now Ministerio de la mujer y derechos humanos) to take overall command of public-sector social policy, under the budgetary control of the Ministry of Finance; institutional innovations including the creation of a national roundtable (Mesa de concertacion en lucha contra la pobreza) as a mechanism to engage civil society in designing social policy and a national agreement (Acuerdo nacional) proposed by the President and supported by civil society representatives that included the reformulation of social policy, a decentralization process and local councils with plans and monitoring mechanisms to keep track of the social programmes implemented. Those programmes had children and youth as top priority groups, and investment in human capital as a key priority (for details, see Annex 3.3). Near the end of the period, a conditional cash transfer (CCT) programme (Juntos) was implemented, focusing on a few districts in the poorest regions of Peru. Garcia’s 2006–2010 government kept some social programmes essentially unchanged while giving others a more market orientation, in particular those in education and health (see Annex 3.2). Political and social pressures helped protect some emblematic food and nutrition programmes (e.g. Glass of Milk). Some long-term programmes (e.g. Parsalud and Huascaran) were retained with diminished budgets and goals, while coverage of the Juntos programme (described below) was extended into most of the poorest districts of the country. Among new programmes, Crecer was the most important. The urgency of reforming the existing social programmes to more clearly define their goals and objectives, to merge them and to identify one public institution responsible for overall social policy from coordination to monitoring and evaluation was widely noted (Vasquez, 2006: 14). In 2007 a social policy reform decree was issued to provide a framework for modernizing management of the programmes (Apoyo, 2009: 24). Some young policy makers and advisers directed their efforts to the monitoring and impact evaluation of social policy, designing new methodologies. Despite government efforts to move

108

Social protection, growth and employment

away from the mostly ‘welfare’ orientation of the 1990s, Peru’s social policy generally retained this perspective, although it now incorporated more market-oriented programmes. The overall effects of recent social policy remain to be assessed.

Resources and managerial issues Total social expenditure increased continuously over the decade, more than doubling for an average growth rate of over 7 percent per year (Table 3.1); its share of overall government expenditure, after dropping sharply from 54 to 47 percent between 1999 and 2001 during the economic and political crisis, has since then fluctuated between 46 and 50 percent, ending the decade at 46 percent in 2010, 8 percentage points below its 1999 level. Around 30 percent of total social expenditure was on administrative costs, with just 70 percent channelled to current and capital expenditures. The administrative share was larger during the Toledo administration (above 35 percent), before falling back to around 26 percent in 2009–2010. Whether either the striking increase in the share of administrative costs over 1999–2001 or the significant decline after 2005 reflect the changes of government at those times is a matter of interest. Around 40 percent (reaching 44 percent in 2010) of total social spending excluding administrative costs went to basic services (pre-school and elementary education, basic health, basic food and nutrition, water and sanitation), while over 50 percent (and over 60 percent until 2005) went to ‘non-basic’ goods and services including secondary and high school education, colleges, infrastructure, electricity, roads etc. Given its rapid increase in absolute terms, social expenditure has also grown sharply per capita, by nearly 6 percent per year and about 83 percent over the decade. This decade saw general government expenditure rise significantly from 15.6 percent of GDP in 1999 to 19.8 percent in 2010 (Table 3.2), while the share of GDP going to social expenditures stayed around 9 percent, with only a marginal net increase over the decade. There was a marked increase in 2009 (from 8.75 percent of GDP the year before to 10 percent), as these expenditures rose sharply and GDP was virtually stagnant (due to the impact of the global financial and economic crisis). Social expenditure was scaled up despite the economic difficulties, and perhaps partly as a response to them.7 A major shift in the composition of social spending saw the rise of health and sanitation from 26.6 to 45.7 percent of the total, reflecting the innovations described in Section 3 below. Between 1999 and 2005 more than 50 percent of direct social spending went to physical infrastructure and equipment projects for initial and elementary education, followed by basic health projects (around 30 percent), while just around 15 percent were for basic social protection and prevision8 (Table 3.3). After 2005 this pattern changed drastically, as the shares for basic (initial and elementary) education and basic social protection programmes were reduced to 40 percent and 14 percent, respectively, by 2010, in favour of a large increase in resources for basic health programmes (to 45 percent by 2010). The supplementary goods and services also had education and health as their main social programmes (both with more than 50 percent of the

Peru

109

Table 3.1: Government, Social and Basic Social Expenditure,1 1999–2010 Type of Expenditure A. General Government Expenditures (Millions of PEN, 2009)

(B/A)*100%

1999

2000

2001

2002

2003

2004

35,183

39,611

46,249

47,284

51,189

54,286

54.3

50.4

46.8

49.5

47.0

48.2

19,104

19,962

21,618

23,389

24,046

26,162

71.2

68.3

63.5

63.4

62.9

1.1 Basic social expenditures

(38.8)

(43.7)

(41.4)

(37.8)

(38.2)

(39.3)

1.2 Supplementary social expenditures5

(61.2)

(56.3)

(58.6)

(62.2)

(61.8)

(60.7)

B. Total Social Expenditure, (Millions of PEN, 2009) 1. Non previsional social expenditures3 4

28.8

31.7

36.5

36.6

37.1

C. General Government Expenditure per capita (Thousands of PEN, 2009)

1.400

1.553

1.787

1.802

1.924

2.014

D. Total Social Expenditure per capita (Thousands of PEN, 2009)

0.760

0.783

0.837

0.891

0.904

0.970

E. Basic Social Expenditure per capita (Thousands of PEN, 2009)

0.210

0.234

0.220

0.214

0.217

0.240

3.38

3.49

3.51

3.52

3.48

3.41

2. Previsional expenditures

Exchange Rate (PEN /US$1)

Type of Expenditure A. General Government Expenditures (Millions of PEN, 2009)

(B/A)*100% B. Total Social Expenditure, (Millions of PEN, 2009)

2005

2006

2007

2008

2009

2010

57,457

61,307

64,089

71,768

79,382

86,695

49.1

48.7

50.2

46.6

48.1

46.3

28,240

29,845

32,181

33,425

38,217

40,143

62.7

65.1

68.0

71.5

74.5

74.0

1.1 Basic social expenditures4

(39.2)

(40.6)

(42.6)

(45.5)

(43.0)

(44.3)

1.2 Supplementary social expenditures5

(60.8)

(59.4)

(57.4)

(54.5)

(57.0)

(55.7)

37.3

34.9

32.0

28.5

25.5

26.0

C. General Government Expenditure per capita (Thousands of PEN, 2009)

2.104

2.218

2.292

2.538

2.777

3.000

D. Total Social Expenditure per capita (Thousands of PEN, 2009)

1.034

1.080

1.151

1.182

1.337

1.389

E. Basic Social Expenditure per capita (Thousands of PEN, 2009)

0.254

0.285

0.334

0.385

0.428

0.455

3.30

3.27

3.13

2.92

3.01

2.83

1. Non previsional social expenditures3

2. Previsional expenditures6

Exchange Rate (PEN /US$1)

Source: author’s elaboration with data from MEF—Estadísticas de Gasto Social (1999–2010) Notes: 1 MEF-DGAES taxonomy of the social expenditure; 2 Includes all national, regional and local, capital and current expenditures; 3 Includes capital and current expenditures; 4 Defined, according to the Oslo Consensus, as: basic education (pre-school and elementary), basic health, food and nutrition, and water and sanitation; 5 Includes activities and social projects not included as part of the basic social expenditure (e.g. secondary and high school education, colleges, social and productive infrastructure, rural electricity, rural roads etc.); 6 Includes capital and current expenditures associated with the work of the social programmes.

110

Social protection, growth and employment

Table 3.2: Total Government Expenditure and Social Expenditure,1 1999–2010 (current prices) 1999 A. General Government

(B/A)*100%

27,254

2000 31,837

2001 37,906

2002 38,829

2003 42,986

2004 47,256

54.3

50.4

46.8

49.5

47.0

48.2

14,798

16,044

17,751

19,207

20,192

22,774

174,422

186,141

189,213

199,650

213,425

237,902

A/C (%)4

15.63

17.10

20.03

19.45

20.14

19.86

B/C (%)

8.48

8.62

9.38

9.62

9.46

9.57

B. Total Social Expenditure C. GDP (current LCU) in million PEN3 5

2005 A. General Government

2006

2007

2008

2009

2010

50,826

55,316

58,855

69,722

79,382

88,021

49.1

48.7

50.2

46.6

48.1

46.3

24,980

26,928

29,553

32,471

38,217

40,757

261,653

302,255

335,529

371,073

382,318

444,460

A/C (%)4

19.42

18.30

17.54

18.79

20.76

19.80

B/C (%)5

9.55

8.91

8.81

8.75

10.00

9.17

(B/A)*100% B. Total Social Expenditure C. GDP (current LCU) in million PEN3

Source: author’s elaboration with data from MEF—Estadísticas de Gasto Social (1999–2010) Notes: 1 MEF-DGAES taxonomy of the social expenditure; 2 Includes all national, regional and local expenditures (Millions of Nuevos Soles (PEN) and percent); 3 World Bank data; 4 Author’s calculations; 5 Author’s calculations.

Table 3.3: Basic Social Expenditures by Type of Expenditure, 1999–2010 (PEN millions, 2009 = 100.0 and percentage shares) Type of Expenditure

1999

2000

2001

2002

2003

2004

Education & Culture

59.4

51.1

48.8

55.2

56.9

53.2

Social Protection & Prevision

14.0

17.5

13.6

12.2

14.5

16.9

Health & Sanitation TOTAL

Type of Expenditure

26.6

31.4

37.6

32.6

28.6

29.9

5,281

5,961

5,698

5,611

5,770

6,461

2005

2006

2007

2008

2009

2010

Education & Culture

53.2

49.5

45.9

47.8

43.2

40.5

Social Protection & Prevision

14.6

13.0

16.6

7.4

13.3

13.8

Health & Sanitation TOTAL

32.2

37.6

37.5

44.8

43.5

45.7

6,932

7,885

9,333

10,872

12,235

13,152

Source: author’s elaboration with data from MEF—Estadísticas de Gasto Social (1999–2010) Notes: Includes current expenditure

Peru

111

available budget), although these education programmes increased their share to 48 percent in 2005 and then fell to 43 percent by 2010, while these health programmes kept their share around 20 percent until 2005, then fell to 18 percent in 2010. The already low share for programmes for social protection and agriculture declined further, while expenditure for transportation and housing social programmes increased. These changes in social expenditure allocation patterns were associated with the different perceptions and priorities of the Toledo and Garcia governments, with the former assigning higher priority to education, and the latter to health programmes.9

Target populations for Peru’s anti-poverty programmes and success in reaching them — some hypotheses As discussed above, in the early 1990s the main goal was to mitigate the expected short-term effects of stabilization and the medium-term effects of growth-oriented macroeconomic policies on poor and extremely poor people. Later those objectives were broadened. Because of management and coordination problems in the social policy area, available data from the programmes themselves do not provide a reliable picture of who the beneficiary families are, but over the past decade the statistically representative household surveys from the National Survey of Households (ENAHO) do permit a reasonable description. Although an individual and/or family may participate (legally or not) in more than one social programme, for simplicity we consider here only the participation of the head of household in the social programme that he/she considers the most important, in the expectation that these figures will tell a broadly representative story.10 As of 2000, beneficiary families were concentrated in two programmes: School Breakfast (44 percent of all beneficiary families) and Glass of Milk (42 percent), with People’s Dining Room also important (11 percent). Similar shares of the families were located in rural and in urban areas, and around 40 percent were non-poor. The heads of households were mostly men (70 percent), typically had elementary or high school education, and nearly all (90 percent) worked in an enterprise with 10 or fewer workers. Some notable changes occurred over the decade: by 2005 and 2009 Glass of Milk became the most important programme for around 50 percent of families, followed at a distance by School Breakfast (27 percent) and People’s Dining Room. By 2005 the percentage of beneficiary families that were in rural areas had risen to 61 percent (from 53 percent in 2000); by 2009 this share had dropped to just under 52 percent. An additional key piece of information that remains to be reliably estimated is the share of all poor families that were beneficiaries of the social programmes. The persistent significant participation of non-poor in these programmes that target poor people may jeopardize their effectiveness (especially if those non-poor are far above the poverty line); participation of non-poor families decreased from 41 percent in 2000 to 26 percent in 2005 but then rose again to 40 percent in 2009. The share of female family heads rose from 26.3 percent in 2000 to over 32 percent in 2005 and 2009. Most of the heads of household had elementary to high school education (and more with higher levels in 2009) and continued working mostly in microenterprises (of two to 10 workers). A logit regression shows that over this decade the main determinants of a family’s participation in a social programme are the number of dependents (positive), poverty

112

Social protection, growth and employment

status (positive) and urban residence (negative); other variables with an impact are gender of head of household (more likely if female) and age (more likely if younger — see Annex 3.4). More complete information and analysis of these social programmes is required; specifically how well they are targeted to nutritional status, height and weight and to certain age groups (e.g. infants, elderly people).

Table 3.4: Head of Household by Social Programme and Characteristics, 2000, 2005 and 2009 (percentage) 2000

2005

2009

51.1

45.5

Social Programme Glass of Milk

41.8

Glass of Milk

Wawa Wasi

0.3

People’s Dining Room

12.2

9.5

Children’s Dining Room

0.2

School Breakfast

26.0

28.3

2.1

3.0

People’s Dining Room Food for Working

11.0

School Lunch

0.2

Baby Porridge

6.6

Family Basket

0.3

Family Basket

0.2

1.5

Food Supplement for at-risk people

0.6

Others

8.4

5.7

Tuberculosis Programme

0.0

Infant Nutrition

0.2

School Breakfast

44.4 100.0

100.0

60.5

52.1

School Dining

1.2

Total

100.0

Total

Rural

53.0

Rural

Location Urban

47.0

Urban

39.5

47.9

Total

100.0

Total

100.0

Gender Male

73.7

Male

67.8

67.4

Female

26.3

Female

32.2

32.6

100.0

14.1

11.4

Total

100.0

Total Education Level

No Level and Initial

14.2

No Level and Initial

Elementary

47.8

Elementary

56.3

47.6

Middle and High

31.7

Middle and High

24.4

31.6

Technical

1.9

Technical

2.9

5.7

University

4.5

University

2.3

3.7

100.0

Total

100.0

Total

Source: author’s elaboration, based on the ENAHO for 2000, 2005 and 2009.

Peru 2000

2005

2009

113

Firm Size 1

22.1

17.2

17.9

2 to 10

66.2

2 to 10

73.4

67.2

11 to 50

5.1

11 to 50

2.6

5.4

6.7

More than 50

More than 50 Total

100.0

Total

6.8

9.5

100.0

Poverty Status Extremely poor

22.4

Extremely poor

36.4

26.2

Non-extremely poor

36.1

Non-extremely poor

37.5

33.7

Non-poor

41.5

Non-poor

26.1

40.1

100.0

Total

100.0

Total

Source: author’s elaboration, based on the ENAHO for 2000, 2005 and 2009.

Evaluation of social policies — main issues During the 2000s, the social programmes mentioned above received resources and mandates to provide specific goods or services, with the programme staff usually defining their specific target group (e.g. women, children up to six years old, children up to 12 years old, extremely poor households, single mothers). Serious questions have been posed about the programmes’ effectiveness—from design and resources allocated to implementation and evaluation. Vasquez (2006) claims that, despite the continuous increase in social expenditure, these efforts did not significantly reduce poverty rates; he attributes this to leakage (where beneficiaries include people not in the target population) and under-coverage (where members of the targeted group do not receive benefits) in all of the social programmes.11 Analysing four nutritional programmes, Monge et al. (2009) argue that the fall in poverty rates cannot be attributed to improvements in targeting, since leakage and under-coverage problems have persisted at high rates. An additional concern is the geographical allocation of the expenditures which, as Lavado (2007) suggests, sometimes favours non-poor regions and locations, even when ostensibly targeting poor people. Issues of information, corruption and political clientelism have also generated problems of identification, principal agent and free riders. The multiplicity of social programmes and the lack of coordination among them have also contributed to difficulties in achieving the right outreach. As overall social policy was continuously redefined from the 1990s on, so too were the specific goals, the programmes, the definition of the poor and extremely poor groups, and the evaluation methods. According to Tanaka and Trivelli (2002: 30–31) failure to take adequate account of these changes led some observers to misleading conclusions in the 1990s, and contributed to misallocation of social policy resources.12

114

Social protection, growth and employment

Failures in social policy implementation have been linked to too few trained staff, low human capital and scarce fiscal resources. There has been a generalized weakness in tasks such as adequately defining beneficiary graduation and exit, with resulting potential for social conflicts among previous and new or potential beneficiaries, and continuing or even rising leakage and under-coverage problems. Underlying more immediate problems such as imperfect targeting are important institutional deficiencies in management and administration associated with corruption and the biasing effects of a government’s client groups.13 The decentralization implemented since 2000 did not have the expected results, partially because of planning problems but also because political decentralization has not been adequately linked to economic and human resource improvements in centres outside Lima, as the economic resources continue to be managed centrally. Peruvian authorities have recognized a number of these challenges and the related concerns of civil society, practitioners and policy makers (Controlaría, 2008). In the last few years a number of highly qualified professionals have become involved, helping to improve various aspects of social policy, from design to evaluation. One of the first innovations was the elaboration of the district-level National Poverty Map in 2003 as a tool for identifying priority locations for social policy interventions. Another was the creation of the targeting system Sistema de Focalización de Hogares (Sisfoh) in 2004, designed to improve the beneficiary selection process of social programmes and to deal with targeting issues (Moron, 2009). New methodologies to evaluate social programmes have also been promoted.

Selected social policies Some social programmes aim to meet just one need and/or deliver one type of good or service, while others provide money or a range of goods and services. The programmes selected to illustrate each of these categories have been selected on the basis of three main criteria: budget size, number of beneficiaries, and potential employment effects.14 They are: • those providing just one good or service: ○○ Food and nutrition social programmes—Glass of Milk; ○○ Education—Literacy programme; ○○ Health—Integrated Health System (Sistema Integral de Salud—SIS); ○○ Infrastructure—Provias; and ○○ Temporary work — Urban To-Work (A trabajar urbano — ATU) and Constructing Peru (Construyendo Peru); • those providing more than one good or service, including cash transfers: ○○ CCT— Juntos; and ○○ Rural production support—Foncodes; and • private social initiatives, including: ○○ Productive projects for small producers—Fondoempleo; and ○○ Development projects for extremely poor people in rural areas—Sembrando.

Peru

115

The Glass of Milk programme Created in 1985 as one of the first food and nutrition programmes in Peru, Glass of Milk delivers one daily portion of food to women and children who are pre-identified as vulnerable to food insecurity. It targets children under six years of age, pregnant women, and women who are breastfeeding. The Controlaría (2011: 59–60) reports that in 2010 the programme’s budget was PEN363 million (about US$128 million), with 2.38 million reported beneficiaries, of whom 68.3 percent were children aged six or under, lactating mothers and invalids; the others were older children, elderly people and people with tuberculosis. From early on, this programme was designed to be run through civil society organizations (organizaciones de base —OBs), created by the Constitution under the Fujimori government of the early 1990s, with a clear gender orientation and mothers preparing and delivering the food. It is funded by the central government and international cooperation. First implemented in poor provincial municipalities around the country, it was soon also installed in Lima, under the argument that food security was a human right. That argument helped Fujimori to develop close political connections with the beneficiaries (Palma, 2008), generating large popular support for his government. Despite various reforms and in the midst of conflicts with poor non-benefited agents, the participation of OBs has continued up to the present. The involvement of local (district) authorities was first reduced (in the late 1990s) then increased again (2001). The programme is currently under municipal responsibility, with two complementary implementing agencies: the local administrative committee (with representatives elected among the mothers of OBs) and the municipal committee, each with its own specific but separate and complementary functions in the areas of selecting beneficiaries, management and control of the financial resources issued by the Ministry of Finance and buying food (from selection and negotiation with suppliers to follow-up, supervision and evaluation). These managerial functions have been vulnerable to inefficiency and corruption. Suarez (2003) identified problems of a lack of oversight plaguing both agencies. Given the criticisms of the programme and its high level of resources, it has been the subject of lively debate among policy makers and political leaders. Previous beneficiaries and the leaders among them persistently oppose graduating from the programme and paving the way for new participants, while protests by current beneficiaries prevent the government from making more than minor changes to its management and supervision. Poor monitoring and ex post evaluation of the programme have also probably contributed to deficient outcomes. Portocarrero et al. (2010) note that the programme has relied heavily on the close ties between the mothers in charge and the government authorities, with the OB Coordinator having the greatest managerial control. Instead of opting for a true communal basis, the mothers in charge have preferred to handle cooking and delivery individually, charging the beneficiaries a (low) fee for these services. These monitoring deficiencies and managerial patterns may explain the programme’s drift away from its original goal of serving only women and children suffering from food insecurity. Positive side effects, however, have been the high

116

Social protection, growth and employment

level of participation of poor women and the economic and political empowerment of at least the leaders among them, although most of them are not from the poorest quintiles, who tend to avoid such responsibilities because of lack of time or interest. This pattern helps to explain the stiff resistance of current leaders to the implementation of graduation mechanisms. To what degree has this programme reached poor and extremely poor households? Various studies have pointed out that, despite the expansion of its coverage, inadequate targeting has been a main weakness. Vasquez (2006: 15–16), using data from the 2004 ENAHO, noted that in that year more than 1 million people benefited through leakage, most of them in Lima and Tacna, regions with low rates of malnutrition, while an estimated nearly 4 million members of the target group were not reached. He attributed under-coverage in most regions more to a lack of resources and of adequate data at the municipality level than to mismanagement. Using the same 2004 survey, Portocarrero et al. (2010: 316–322) report that Glass of Milk had comparatively good outreach among the poorest households; quintile 1 and 2 beneficiaries constituted almost 60 percent of the total, while 25 percent came from quintiles 4 and 5. By 2008 these shares were 57 percent and 30 percent, respectively. In perhaps the most complete analysis of this programme as of its publication, Stifel and Alderman (2006) also judged that it was relatively well targeted to poor households and those with low nutritional status, primarily due to targeting poor districts and because poor households received larger in-kind transfers due to intra-district targeting. Taking the analysis one step farther, however, they find no clear impact on nutritional status, since the transfers of milk and milk substitutes are inframarginal for about half of the recipient households. Although this conclusion does not imply that the programme has no or a low payoff, it does highlight that to achieve a major nutritional impact requires even more precise targeting than simply getting milk to families who fall under a specified monetary poverty line, and that effective monitoring and evaluation is even more complicated than it initially appears to be. On the other hand, the comptroller’s office (Controlaría General de la Republica, 2011) and Professor Enrique Vasquez of Lima’s Universidad del Pacifico (Páez, 2010) have negative views on the quality of targeting. Given this programme’s importance, further analysis is clearly needed, both to assess the social implications of leakage and of under-coverage when they appear, and to assess their determinants. As noted earlier, leakage is less serious when the beneficiary is not far above the poverty line, and under-coverage is most severe when a nonbeneficiary is far below the poverty line. Regarding determinants of leakage, Yokoyama (2007) reports that when mothers’ committees in charge of the programme were monitored from below by the beneficiaries, leakage was only half as great as when there was no monitoring, whereas when the monitoring was top-down (by local government), leakage was the same as in the absence of monitoring. The programme generates no waged employment. The mothers prepare the food for the beneficiaries’ breakfasts within their own homes in the mornings, so a modest amount of productive employment is involved. Indirect effects on employment and income may, however, arise within the supplier firms, if they hire additional labour to produce the goods required for the programme.

Peru

117

Integrated Health System In 2002, following complaints about leakage and under-coverage in two previous socialized health programmes—for schoolchildren and for mothers and infants, respectively—those programmes were merged into the new SIS programme, created as a decentralized public office under the Ministry of Health; in 2008, in the government’s modernization process, SIS became an autonomous public agency. Since its inception, the programme has relied on existing public hospitals and the public health centre system to provide services to the intended beneficiaries; in return these hospitals were to be paid by the SIS for the professional care and medicines provided. The ambitious declared goal of the SIS was to provide fully or partially subsidized health insurance to the poor and extremely poor population and those with no resources to cover insurance payments. Currently there are three programme plans: one is totally free for the poorest families; a second covers independent workers (or microentrepreneurs) who can gain personal and family access to SIS by paying a monthly fee of US$8–15, which allows them larger coverage than that of the first plan; and the third is for workers of formal microenterprises (and their families), who may be enrolled by the registered microenterprise owner, who—as employer— has to pay a subsidized monthly fee (around US$5 per worker). The latter two programmes were initially launched in a few cities and have since been expanding throughout the country. SIS’s declared goals of reducing chronic mother and infant undernourishment and, more generally, individual health risks within the poorer segments of the population have been at least partially achieved over the course of the decade. The system has continuously broadened its coverage to include more complicated health treatments and services. Its personnel are in charge of enrolling beneficiaries, verifying availability of services, supervising medical attention, verifying public satisfaction and processing the payments to the health centres providing the services. Informational campaigns and the simplicity of enrolment in the programme (with official identification and a form declaring incomes and poverty status) have multiplied the number of beneficiaries quickly, at a rate that the budget has not been able to match. By 2010, the programme reported over 12 million insured people, or more than a third of the country’s population. According to official SIS statistics, one third of the beneficiaries were children under 10 years, 28 percent were young people of 10–19 years, 33 percent were working-age people (20–59 years old), and a little under 7 percent were elderly people. Two thirds of the beneficiaries came from the poorest 40 percent of the country’s population; in terms of location, 33 percent lived on the coast, 43 percent in the Andes, and the rest in the Peruvian Amazon. More than 96 percent were covered by the SIS Plan 1, while just 4 percent were under Plans 2 and 3. In 2010, the reported value of SIS transfers (without administrative costs) was almost PEN450 million (more than US$160 million). The client population complains about limited and slow medical attention, while the public hospital administrators and professionals complain about payment delays and underpayments from SIS. The combination of an enormous increase in SIS users (a doubling between 2006

118

Social protection, growth and employment

and 2008, for example), no additional budget for investment in physical infrastructure and only minimal increases in that for medical professionals and medicine, along with serious underpayment by beneficiaries of their contributions, has in effect created an excess demand for the public health systemâ&#x20AC;&#x2122;s services. The immediate consequence has been a de facto lack of coverage for a large proportion of the target population (Seinfield and La Serna, 2007), in particular the poorest families, many of whom are located in small towns and other places not yet included in the programme. The quality of health services has also fallen, in terms of both care and available medicines. These problems are the natural result of a sudden surge in demand which would be hard to cope with under the best of conditions but which is aggravated by underlying inefficiencies such as inadequate monitoring, deficient information systems and the weak capabilities of managers. Peruâ&#x20AC;&#x2122;s experience with SIS is similar in many ways to that of other health programmes. The demand for subsidized services is very high, the intent is laudable, but the resources are inadequate, and not enough planning has preceded implementation. Given the inevitability of a gestation period during which a system such as this increases the quantity and quality of its services, it is important to learn what could have been done better and how to proceed now. Overloading the scarce health professionals can clearly be counterproductive in the medium term. There may be trade-offs between targeting the poorest first, keeping costs down and keeping professionals satisfied. Fortunately Peru, with its rapid increase in public revenues from mineral-based growth, has more options available than many other countries.

Provias The Proyecto especial de infraestructura de transporte nacional (Provias Nacional â&#x20AC;&#x201D;PVN) was created in 2002 under the Ministry of Transport and Communication (MTC) as successor to the previous Programa especial de rehabilitacion de infraestructura de transportes (PERT). It was given technical, administrative and financial autonomy as a large temporary programme in charge (from design to execution) of transport infrastructure projects, including construction, improvements, rehabilitation and maintenance as well as planning, management and control of activities and resources. Two parallel programmes, Provias Departamental and Provias Rural, were created to handle specific local transport projects; in 2006 both were merged with the new Provias Descentralizado (PVD) to work with regional governments. All of these programmes have worked with private-sector firms selected through a bidding process; depending on the contract arrangements, private firms may be in charge just of the execution of the project or also of various aspects of its maintenance. By 2008, the overall programme dispensed around US$1 billion (more than PEN3 billion) for its different road projects, with more than 50 percent from PVN, more than 17 percent from PVD, and the rest mainly from regional governments (MTC, 2009: 2). The PVD was partially funding and providing technical assistance (in coordination with regional governments) to regional and decentralized rural transportation. Among sources of financial support were the World Bank and the Inter-American Development Bank (IDB).

Peru

119

The Provias programme’s main social dimension lies in its contribution to low-income rural producers by giving them better roads, easier access to markets and hence better farm gate prices for their products. It also appears to have had important effects on local employment, as expected given the nature of its projects. Although no statistics on the total employment generated have been officially reported or otherwise estimated, it is plausible to expect that such local employment in road maintenance and rehabilitation would have positive reactivating effects on small, often sluggish local economies and their labour markets. Further indirect employment effects may be associated with the raw materials used in the projects, although most of the producer firms are not local but, rather, located in Lima or even abroad. The economic logic of building or enhancing rural roads that help to connect low-income farmers to the market has long been recognized, especially when the cost of construction or improvement is modest and much of the work can be carried out by local workers. The World Bank (2010) tallies the benefits: in 11 years more than 15,000 km of rural roads had been rehabilitated, more than 500 microenterprises created to maintain the roads, about 6000 permanent jobs created and a major reduction of travel times (by 30–60 percent) achieved. Before this project, “low-cost gravel road alternatives were discarded because communities were afraid they would not last” (MTC, 1999: 2). As of 1999 only 28 percent of rural households had access to a road in good condition, and freight transport services were infrequently available, if at all. Subsequent evaluations showed that 78 percent of households surveyed considered the roads to be adequately rehabilitated and maintained. A 2005 evaluation pointed to significant improvements in school enrolment (8 percent), visits to health centres (55 percent), land use for farming (16 percent) and male agricultural salaries (20 percent). Rural entrepreneurship was strengthened. After overcoming initial male resistance, women came to account for 24 percent of the workforce in the microenterprises by 2005. To increase the programme’s benefits, a Local Development Window (LDW) was introduced; this mechanism, managed by non-governmental organizations (NGOs), identifies and prioritizes local productive initiatives. Institutional improvements, increased local participation and better local governance both helped and were encouraged by the project. A learning process was built in. For example, the fact that a measured impact on poverty was delayed by several years after roads were rehabilitated provided part of the incentive for the LDW (World Bank, 2010: 4). Involving women in the design of projects was critical to meeting their needs as well as men’s. Participatory assessments resulted in the construction of footpaths, preferred by women because they are the safest way to take animals to pasture and to collect firewood and water. The project provided grants for women to start microenterprises. According to the World Bank (ibid.: 5), the programme has become a benchmark for best practice in other rural transport operations in Latin America. This suggests that it can be seen as a step towards narrowing the gap between the rural development experience of a Latin country and those of East Asian success stories such as Taiwan in the early decades of their rapid and equitable development, when public investment in infrastructure was strong and rural synergies between agricultural and other activities were prominent.

120

Social protection, growth and employment

Although its economic logic is impressive, implementation problems can always hold down the payoff to programmes like this one. Academic assessments, generally supporting the positive interpretation of World Bank documents, suggest that local ‘capture’ and corruption threats were controlled reasonably well. Escobal (2005), based on a variety of methodologies, concludes that there are important complementarities in rural infrastructure development, such that even if a given type (roads, electricity, water) is subject to diminishing returns, this effect can be overcome by advancing simultaneously on enough of these fronts. Efficiency and equity gains may not go together, especially at the start, because the better-off families have a stronger private asset base that allows them to make better use of the public investments. Valdivia (2010), using a carefully designed quasi-experimental approach, confirms that motorized rural roads alter employment patterns, raise investment in education and health and bring greater income and other benefits when complemented by investments in other areas. Income effects were not significant on average, which is perhaps not surprising, since the survey from which the data were obtained was undertaken only two years after the intervention began (ibid.: 22). Such effects were strong in villages with pre-existing endowments of key productive infrastructure and may well have become generalized over time. Most of the positive results relate to interventions in motorized roads, although the tracks that help women enable a different and less easily quantifiable sort of benefit. Escobal and Roman’s (2008) econometric analysis of the impact of rural roads implies an annual income gain for beneficiaries on motorized roads of US$121—or more than 35 percent of prior income. Residents near non-motorized roads did not show a significant income gain, attributed by the authors to their not having access to as many other public assets as the former group. Non-agricultural wage incomes rose by US$114 and US$69 for beneficiaries on motorized and non-motorized roads, respectively, while the latter group saw a reduction of US$97 in non-agricultural self-employment income (the former group saw no significant change); presumably this loss occurs due to the increased importation of goods that displaces local production. The roads did not significantly affect consumption of either group. Motorized road beneficiaries invested heavily in livestock, to the tune of 56 percent of annual income. No such effect was observed for families on non-motorized roads. One challenge to measuring the programme’s effects is to adequately assess the benefits accruing to women from the non-motorized roads, given that many or most of them do not take monetary form.15 Other challenges are suggested by the fact that some studies of Peru’s investments in rural roads have found no general pattern of income gains, although favoured beneficiaries near motorized roads have clearly benefited. Leaving aside the possibility of measurement error, a plausible hypothesis is that losses of some unspecified sort could be offsetting more easily observable gains or that some groups gain but only after a considerable lag. Attention must be directed to understanding potential losses as well as the time trajectory of both benefits and losses, the latter of which call for careful follow-up analysis. A final desirable form of analysis around a project such as this one, where the initial evidence is encouraging, is a look at how big such an intervention could become. Such an

Peru

121

assessment entails both taking account of the benefits and costs observed thus far and then assessing how much more rural infrastructure investment could be undertaken with comparably positive effects. The benefits would normally diminish at some point either because the payoff to the new infrastructure diminishes or because the labour absorption becomes less important as the rural labour market tightens up, or for other reasons.

Special labour programmes: ATU and Constructing Peru Various temporary work programmes have been created over the last couple of decades. Two of the more important, judged by their budget, coverage and employment effects, are ATU and Constructing Peru, both under the supervision of the Ministry of Labour. ATU started in 2002 to mitigate the loss of employment during the ongoing recession by generating temporary jobs for heads of households from the poorest urban localities to work on small local economic, social or environmental protection projects such as water and sewage systems, sidewalks, pipes and walls. Projects were carried out by institutions (public, private or civic) willing to contribute with some resources to the projects and selected through an open competition, all under the supervision of the Ministry of Labour (Lizarzaburu, 2007). The selected organizations hired unskilled workers for around four months at below minimum wage, selecting workers who had just lost their jobs, on the assumption that they were the most affected by the crisis. Candidates just needed to present official identification and proof of having one unemployed dependent under 18 years of age. The daily wage of around US$5 (or around US$100 per month) was set well below the minimum legal salary (around US$250 per month) to avoid perverse effects on workersâ&#x20AC;&#x2122; decisions and to avoid unfair competition in regular labour markets. This arrangement ensures that participants are those with the least attractive options in the local labour market. Official estimates indicated that ATU created 230,000 temporary jobs between 2002 and 2006 while implementing more than 500,000 local projects. Lizarzaburu (2007) identifies a significant change in the programmeâ&#x20AC;&#x2122;s execution by late 2003, stressing its initial success, measured by the number of competitions and projects implemented and also by the progress in establishing solid managerial and application mechanisms. After 2003 the ATU became more permanent in character; with the crisis over, the budget was considerably reduced, there was increased rotation of responsible authorities, and the focus shifted to physical investment projects in marginal urban and poor rural areas, but with reduced hiring of unskilled workers because of the altered nature of the projects. The main consequences of these changes, evident by late 2006, were lower employment generation and lower income transfers towards the original target population. Chacaltana (2003) criticized this programme on the grounds that its benefits did not match its costs, since the fiscal expenditure involved was greater than the final income gained by the beneficiary families. He noted the low unemployment rate among poor people, most of whom were engaged in low-income informal activities. His Propensity Score Matching analysis, comparing average incomes between otherwise similar beneficiary and non-beneficiary workers, in

122

Social protection, growth and employment

2002 found the average incomes of the former (PEN300) exceeded those of the latter (PEN227) by 32 percent. The author’s conclusion that these levels were ‘similar’ is an extreme interpretation, since for families in poverty or extreme poverty such a difference could be quite significant. His analysis also raises a more general methodological question of how such ‘benefits’ should be compared to the programme’s fiscal costs, and to what reference points its benefit–cost ratio should be compared — for example, those of other programmes and policies that may reduce poverty or income insecurity. In 2007 this programme was transformed into Constructing Peru. While its main goal remained the generation of temporary employment, mainly for those in poor and extremely poor conditions, there was now increasing attention to the development of skills of the formerly unemployed beneficiaries. As before, the main mechanism has been a competition among local organizations for local, labour-intensive projects in construction and services. Individuals, private organizations and local or regional public institutions have become the main counterparts. Between 2006 and 2010 the programme generated more than 300,000 temporary jobs, with almost 100,000 of them located in Lima (Jaramillo et al., 2009: 90).

The Juntos conditional cash transfer programme Inspired by international programmes such as Progresa—Oportunidades of Mexico, the Toledo administration created Juntos in mid-2005, with the declared objective of reducing poverty and breaking its intergenerational transmission cycle. Its main immediate goals were to improve the access of poor families to quality education, health and nutrition and in the process to reduce the income gap between poor households and the better off, to further gender equality, to respect cultural diversity and to ensure the right to official identity for all. Initially under the PCM, Juntos soon gained administrative and financial autonomy; even its monitoring evolved toward external and internal evaluation systems. Unlike the programmes discussed above, Juntos is a typical CCT programme that delivers monetary resources to low-income families with a view to increasing their basic nutrition, education and health under either voluntary or enforced criteria (for example, no money is transferred until the beneficiary proves that they have met minimum requirements of their children’s school attendance and health checks). Its main target population is poor households where at least one member is a pregnant woman, a widowed parent, an elderly person or the tutor of a child under 14 years of age. Juntos provides PEN100 (around US$35) per month to women from households selected in part using the government’s targeting system (SISFHO). In return, the women participate in health programmes—nutrition (i.e. it is mandatory that children of pregnant women and others under five years of age attend a health centre and accept the assistance offered), education (i.e. school enrolment and attendance of children aged six to 14) and identity (i.e. everyone, including children, gets their official national identification document). Juntos started in one of the poorest Peruvian regions (Ayacucho) and in one of its poorest districts. By late 2010 almost 500,000 households had been reached, mostly in poor and extremely poor localities.

Peru

123

The effectiveness of the programme’s targeting and use of resources has received serious evaluation. Between 2005 and 2010, when limited publicly available data hampered independent assessment, national experts raised various concerns, including leakage and under-coverage. In 2008 the programme was redesigned to include two new components: a support group to advise and provide technical support, and an analytical effort to be carried out by research centres, including the IDB and the World Bank. In its recent (second) evaluation16 (Perova and Vakis, 2011), the World Bank concludes that it has been broadly successful and, as in other countries, has strengthened the social protection system in Peru. It has had “a moderate impact in reducing poverty and increasing monetary measures of both income and consumption” (ibid.: executive summary), has increased the utilization of health services by both children and women and has improved the nutritional intake of households. As in other countries where primary school attendance is high, the main impact on school attendance occurs at transition points (entry and graduation from primary). The majority of the indicators are increasing with the length of the programme. Despite these positive effects, no impacts are found on final outcome indicators such as malnutrition or anaemia, consistent again with international experience, and possibly pointing to the need for programmes such as this one to be complemented by an adequate supply of health services and interventions to improve quality. There has been no evidence of unintended and undesired effects such as a reduction in adult work effort, increases in fertility or in consumption of alcohol. The authors consider Juntos to be superior to other Peruvian social programmes, since it has already succeeded in reducing poverty (measured in monetary terms) and has had positive impacts on health, and there are other cumulative effects to be observed over time. Though generally positive, these results raise a number of important issues. If nutritional intake is in fact improved, why does this not show up in indicators of malnutrition? Is there a natural time lag involved or something potentially more serious to the programme’s ultimate success? Does the ‘moderate’ impact on income, consumption and poverty reflect simply the modest size of the transfers (Lustig, 2011) or is there a degree of inefficiency involved? What can be done to counteract the growth slowdown noted by Perova and Vakis (2011), while the programme still only reaches part of the poor population? Two main challenges confront Juntos: in the short term, to continue to reduce poverty among poor households; and in the medium to long term, to increase investment in human capital to help break the intergenerational transmission of poverty.

The experience of Foncodes Foncodes was created in 1991 by the Fujimori government in the aftermath of the 1990 recession and hyperinflation, when poverty incidence exceeded 50 percent of the Peruvian population. After some reformulations Foncodes continues two decades later, now under the Ministry of Women and Social Development (Mimdes). Its broad goals are the social and economic advancement of the poor and extremely poor population through community-based activities designed to develop their capabilities and meet their basic needs, using both public (various

124

Social protection, growth and employment

levels of government) and private funding. Public funds include not only those from Mimdes but also from other ministries with similar interests and projects (e.g. housing or education). Additional resources come from soft loan agreements with international organizations (mainly the IDB and also the Japanese Bank for International Cooperation (JBIC), the World Bank and the German bank, KFW). Foncodes promotes four types of interventions and projects aimed at: • developing institutional and personal capabilities, aiming to improve the social capital of the communities through planning and managerial tools for regional governments, ‘executive nuclei’ (see below) and local leaders; • economic development projects, aiming to strengthen local production of staple foods and agroindustry through investments in irrigation, soil preservation, reforestation, commercial infrastructure, small processing centres, small market centres etc.; • social infrastructure projects, mostly small investments (typically around US$35,000) in public schools, public health centres, drainage pipes, drinking water systems, latrines, basic bridges and sidewalks and small electricity networks; and • ‘territorial articulation’ projects, oriented mainly at formulating and coordinating the Foncodes interventions with the public and private sectors, with social programmes and with regional and local governments, to achieve better management of its projects throughout the country. In this regard, an important tool developed over a period of time by Foncodes, in cooperation with the Ministry of Economics and the National Statistics Institute, has been a National Poverty Map to identify poor populations within the smallest geographical division level (the district). Since 2003 it has become a key tool in allocating Foncodes projects and delivering its services to poor households. Since 1996 Foncodes has operated through the ‘executive nucleus’, a representative community-based entity with sole authority to do business with Foncodes. Foncodes coordinates with regional authorities about the needs of poor people, to assess the possibilities for one of its own types of intervention; it then identifies and announces a set of projects for which the executive nucleus calls for local proposals. Foncodes provides the bulk (usually 90 percent) of the funding, while beneficiaries contribute the rest, mainly in the form of community labour and physical space. The executive nucleus selects the raw material providers, organizes community labour participation, supervises project execution and oversees maintenance. Through its local personnel, Foncodes monitors overall project execution and at the end transfers project ownership to the community. Its microcredit programmes, mainly implemented through public financial institutions, receive special attention (see Section below on “Microenterprise and microfinance in Peru: what about social policies?”). Over the last two decades most of the projects approved by Foncodes have been located in the rural Sierra (although in the last few years an increasing number have gone to the poorest areas of Lima and Callao) for child-care services, beginners’ education, public health centres, secondary electricity networks and roads. According to Alcazar and Wachtenheim (2001), more

Peru

125

than 30,000 projects were implemented during 1990–1998 at a cost of over US$1 billion, with around 50 percent for social infrastructure (e.g. educational centres and drainage pipes) and most of the rest for social assistance (e.g. nutrition and health centres) and economic infrastructure (e.g. irrigation and local transportation). Between 1991 and 2011, Foncodes issued around US$2.5 billion to fund 56,800 projects, around 40 percent of them for social infrastructure (e.g. educational centres, latrines), another 40 percent for economic infrastructure (e.g. tourist lodges, crop collection centres) and minor projects in diverse activities (e.g. slaughterhouses, small training centres) and finance (including microcredit). In January 2012, Foncodes was transferred to the new Ministry of Development and Social Inclusion (MIDIS), reflecting the new government’s view on social policy and its role in reducing (mostly rural) poverty. Based on its size and duration, a big demand-driven programme such as Foncodes should have yielded considerable benefits in absolute terms. Its nature, however, makes it hard to assess its positive effects on poor and very poor households. Based on a representative (but now dated) survey, Alcazar and Wachtenheim (2001) concluded that between 1994 and 1999 project success was not guaranteed by community participation and outreach, but that the level of development of the community, its human capital, training programmes linked to the project, and the presence of public institutions in the local community may have contributed to project success. Francke (2001) pointed out that local professionals—working with the executive nucleus—frequently developed clientelistic connections with Foncodes employees, although this does not detract from the projects’ favourable effects. More serious problems have been quality control, maintenance, narrow coverage, lack of training for maintenance etc. Has the Foncodes programme had significant employment effects together with the more direct benefits from the investments undertaken? Neither employment generation nor the promotion of labour-intensive technologies has been an explicit criterion for project selection, but the nature of most of the projects implies a considerable impact on labour demand. Since 1997 most of the Foncodes social projects have been implemented in areas with low or very low open unemployment rates, but, because of the nature of the projects, they do generate direct and indirect employment.17 Francke and Espino (2001) highlight the different potential across types of projects to generate seasonal or temporary employment, identifying those with highest employment generation as including land recovery, construction of small-scale tourist facilities, drainage works, sanitation and small hydroelectric systems. Priority projects in terms of budget and beneficiaries have been small drainage pipe systems, small irrigation systems, and drinkable water, all of which are labour-intensive. Indirect labour effects may come from either inputs required or increases in consumption.

Some private social initiatives Together with the approximately 50 public social programmes in operation as of mid-2011, there have been a number of others designed, supported and funded by the private sector. Going beyond a corporate social responsibility approach, these initiatives focus explicitly on specific dimensions of poverty. Their usual goal is to provide funds and technical assistance for certain

126

Social protection, growth and employment

target groups. The two programmes reviewed here are designed to strengthen the skills of lowskilled workers in previously identified (poor or in-need) areas, with a view to helping them to generate sustained labour incomes. Fondoempleo The National Fund for Labour Training and Employment Promotion (Fondoempleo) is a private autonomous organization in operation since 1998 that supports projects designed to lead productive employment with dignity and higher incomes for currently vulnerable workers by developing and strengthening their capabilities. Funding comes from participating private firms from the same department (state) as a given project. For many years this programme has been overseen by the public sector, despite its private nature. The fund’s financial resources come from the net profits of the firms in the department involved (from large firms, mostly mining) as established by law, plus any funds from new voluntary contributors. Based on these resources, Fondoempleo covers most of the proposed expenses, with the collaborating institution assuming the difference. In 2010, it disbursed PEN80 million (almost US$30 million) in 12 departments. Fondoempleo operates through an annual call for projects in the form of an open public competition throughout the country. Although it usually contracts a third party to manage the competition, it retains responsibility for evaluation, selection, funding and monitoring. Final criteria for selecting a project are the quality of its design and its potential to improve local employment opportunities and productivity. Projects approved vary widely both by activity (e.g. agriculture, livestock, tourism, small-scale textiles) and by beneficiary group (microenterprises, women). Most are executed by small and microenterprises, consulting firms, NGOs or any other public or private institution with more than four years of operation and judged capable of undertaking the proposed project. Because employment generation is the main goal, all the projects approved are labour-intensive and generate labour income directly and sometimes indirectly. The employment effects may be significant within a given locality. No detailed studies of the overall labour effects are available, however. Despite the significant resources involved, Fondoempleo’s regional scope is limited by the fact that funding must come from firms located in the same department as the projects. Sembrando This programme was operated during President Garcia’s second administration (2006–2011) by the NGO Instituto de Trabajo y Familia, formally headed by the President’s wife but fully funded by international cooperation and donations from private individuals. The main declared goal is the social and economic development of extremely poor families in rural areas situated at an altitude of above 2500 metres, by creating the material conditions to break the vicious circles of poverty and exclusion. To this end, the programme aims to make the houses of these families healthier by introducing improved kitchens and ecological latrines, reducing malnutrition of children under five, and raising agricultural productivity through the use of improved seeds and better harvesting techniques. Its basic approach relies on inducing the communities to work with the NGO and involve their representatives in the projects. After a successful contact

Peru

127

between Sembrando and the community, a development baseline and then a multidimensional strategy are set. Sembrando appears to have worked well in relation to its main goals. It has reached an increasing number of beneficiary families (over 10,000 in 2007 and about 28,000 in 2009). The key service has been the improved kitchens and the ecological latrines; others include literacy, advice in obtaining identification documents, and training. As with most other targeting programmes, direct employment generation has been minimal. The results have been appreciated in Peru and abroad though, and Mrs. Garcia has gained international recognition for this work.

The overall impact of Peru’s social programmes Since most social programmes are created to deal with one or another aspect of poverty, it is logical to ask to what degree their success can be measured by an analysis of poverty trends. As expected, the incidence of poverty in Peru (as defined by income) has fluctuated fairly closely with GDP per capita, which bottomed out in the early 1990s and only exceeded its late 1980s average in 2003–2004. But poverty reduction was not an automatic result of rising GDP per capita, and over 2001–2003 poverty reached its highest levels since the early 1990s. It then experienced a continuous drop for the rest of the 2000s, as per capita incomes rose rapidly (Table 3.5). Since 2000, poverty reduction has been more marked in urban areas (with a slightly

Table 3.5: Incidence of Poverty by Area (rural/urban), 1981–2009 (percentage) Year

National

Urban

Rural

1981

51.2*

32.9*

83.6*

1993

53.9*

39.2*

88.2*

1995

45.3

37.4

59.8

1996

44.1

36.9

57.0

1997

47.6

33.7

72.7

1999

48.6

36.1

72.5

2001

54.8

42.0

78.4

2003

54.7

43.1

76.0

2004

48.6

37.1

69.8

2005

48.7

36.8

70.9

2006

44.5

31.2

69.3

2007

39.3

25.7

64.6

2008

36.2

23.5

59.8

2009

34.8

21.1

60.3

* Figures based on the Unsatisfied Basic Need methodology, as reported by the Instituto Nacional de Estadistica e Informatica (INEI) (1994: Table 8). The other years reported poverty line defined in monetary terms. Sources: INEI (1994; 2009).

128

Social protection, growth and employment

larger percentage point drop and a much large percentage fall than in rural areas). Poverty rates in rural areas have been very high and are now about three times the urban level. The degree to which poverty reduction has been a result of social policy is hard to judge. The issue is not whether that policy has been the principal source of poverty reduction –implausible in light of the rapid increase in per capita income and the modest amount of assistance probably getting through to poor people—but, rather, whether or not it has made a significant contribution to the decline. Although Table 3.4 shows a considerable leakage of benefits to nonpoor families, if the reporting is accurate, 60–74 percent (depending on the year) of beneficiary families have been poor. No doubt a number of non-poor beneficiaries in a given year moved in and out of that state or were not far above the poverty line; in either case, their inclusion in a programme is not a bad thing per se, even though it would presumably have been still better had the targeting been more precise. Such complexities show that using a single income-based poverty line to assess trends over time is overly simplistic; a more detailed statistical picture of trends in household economic status could add to our understanding of how socially beneficial the programmes discussed here are. Thus, although the available data suggest a considerable coverage of the set of programmes taken as a whole, they do not provide information that can be applied directly to help us estimate how much poverty reduction the various programmes might achieve. This is especially the case if an income definition of poverty is used and the figures for income do not include the value of the social services received (the usual situation); a basic needs definition may allow a better assessment. Even though there may be no as yet accepted metric for doing so, social protection programmes should also be measured against the criterion of economic security. In this respect, programmes such as Glass of Milk may be expected to score high. Two broad approaches can be used to estimate the aggregate effects of social programmes on poverty (or other relevant variables): (i) econometric analysis in which such outcome variables are regressed against their known or presumed determinants, including the social programmes of interest; and (ii) analysis based on microeconomic estimates of the incidence of the programmes on poor people (or other groups). The former has not been attempted in Peru. Incidence figures such as those presented above can be used in the latter approach. One systematic attempt to gauge the poverty incidence of social protection programmes and of government tax and spending more generally is that of Lustig and collaborators (reported in Lustig, 2011) for five Latin American countries: Peru, Mexico, Bolivia, Brazil and Argentina). The estimates for Peru refer to 2009. There is a striking range of experiences across these countries— for example, overall social spending (defined as cash and in-kind transfers plus benefits from public spending on education and health)18 ranges from 17 percent of GDP in Brazil to 5.2 percent in Peru.19 Direct cash transfers are at 0.5 percent of GDP in Peru, compared with 3.1–5.1 percent in Brazil, Argentina and Bolivia; in these three countries, most of the cash transfers arise from non-contributory pension schemes, which tend to be negatively redistributive but are much less important in Peru. Direct cash transfers reduce extreme poverty greatly in Argentina (from 13.5 to perhaps 5 percent) but only marginally in Peru, even including the important in-kind food transfers.

Peru

129

The estimated income redistribution effect of the budget (taking account also of taxes) also varies widely; the Gini coefficient falls by 13 points in Argentina (from a pre-tax and spending level of 0.48 to a post-tax and spending level of 0.35),20 whereas in Mexico it falls from 0.502 to 0.433 and in Peru from 0.495 to 0.470, substantially the smallest reduction among these five countries. Poverty measured at a US$4 (purchasing power parity) level is estimated to fall almost not at all (from 28.8 to 28.1 percent), and extreme poverty (US$2.5) very modestly from 15.1 to 13.9 percent (ibid.: 26, Table 1). The main reason for this smaller progressive effect on equality in Peru is the lower level of social spending relative to GDP; in fact, given the lower levels of social spending, Peru (and Mexico) get the most redistributive impact per unit of spending (ibid.: 15). Households become net contributors to the fiscal system from the fourth decile up—i.e. the first three deciles receive more in social spending benefits than they pay in taxes. Lustig et al. (2011) see Peru as a prototype case of low government spending which redistributes little in absolute terms but does get a good redistributive effect per dollar spent. In this respect it fits the conventional wisdom with respect to Latin American patterns of redistribution through the budget—little redistribution because revenues and social spending are low.21 This study judges a number of the programmes to be positively redistributive, especially Juntos, the food transfer programmes, the SIS health system, and primary and secondary education spending, whereas ESSALUD (the contributory national insurance coverage that is part of the standard package for formal-sector workers) and pensions work markedly the other way (ibid.; see Table 5e). Note that these estimates are based only on what is classified as ‘social spending’, which does not include programmes such as Provias or the employment schemes discussed here. Recent studies such as that of Lopez-Calva and Lustig (2010) identify a pattern of falling inequality in the majority of Latin American countries, including Peru over the last decade or thereabouts, which they attribute to a combination of decreasing earnings gaps between people with more and people with less education, and an increasing contribution of social spending to the non-labour incomes of the lower deciles. The decreasing impact of educational disparities on labour income is attributed to some combination of falling variance of educational attainment and a reduction in the disequalizing effects of technological change, which was particularly noticeable in the period after trade and economic liberalization. A small impact is attributed to demographic changes, as the difference between the larger families of lower-income households and the smaller ones of higher-income households diminishes. Several continuing measurement challenges mean that the conclusion that inequality has been falling significantly in Latin American countries is not yet completely firm. Of particular significance is the failure of the household surveys that provide the raw material for the estimates of inequality to capture a high share of capital income. Meanwhile the empirical evidence points to an increasing capital share over recent decades.22 Although trends in poverty and income inequality can provide one test of the impact of social spending, the benefits of such spending do not depend per se on whether poverty or inequality are rising or falling. It is highly probable that labour income inequality has been falling, and it seems clear that social spending is having an equalizing impact, some measurement problems in some countries notwithstanding. This is an undisputed benefit, even if overall inequality happens to be rising, an unknown at this point.

130

Social protection, growth and employment

Whatever the degree of success Peru has achieved with its social policy, either in absolute terms or in comparison with other countries, many barriers remain to improving outreach and resource use efficiency. Drawing on the experience to date will be crucial in the design of the future social policy of Peruvian governments, including the one just initiated.

Box 3.1: The new Humala government and its social policy agenda — a quick glace Since July 2011, Peru has had a new government, led by President Humala. In his campaign Humala stressed social inclusion as the centrepiece of his new economic and social reforms and policies. His proposed model of economic growth with social inclusion aims to reduce poverty while sharing the benefits of growth widely across the population. Specific social policies proposed include (Gana Peru, 2011): • doubling the coverage of the Juntos programme; • increasing the minimum wage from PEN600 (more than US$200) to PEN750 (around US$250); • introducing Cuna Más for babies in the poorest districts of Peru; • a pension of PEN250 (almost US$100) for people older than 65 and with no insurance except SIS; • Beca 18, to help young people from public high schools to go to a public or private university to study for a degree in applied science; • Beca de Postgrado, for postgraduates to study abroad at prestigious educational institutions; • strengthening the training of primary and secondary teachers; • efficient supervision of the quality of medicines and promoting generic medicines; • Mobil Assistance for Emergency System (SAMU), well-equipped, in the whole country, with an emphasis on rural areas; • school breakfast and lunch for zero under-nourishment; • extra attention to one-teacher, multi-grade schools; • public utilities (electricity, drinkable water and drainage pipes) for all public schools; • action against drug trafficking; • increasing the number of police and police stations.

Peru

131

Microenterprise and microfinance in Peru: what about social policies? Peruvian microenterprise and microfinance—a brief review As noted in Chapter 1, microfinance is one of the important institutional innovations of recent decades in developing countries. Its aim is to raise productivity and incomes in small, mainly non-agricultural firms and thereby to lower poverty and, in the best of cases, provide an important motor for equitable growth. Peru is often considered to be the country in Latin America where such hopes are most justified, based on two conditions: first, the informal or microenterprise sector is very large both in absolute terms and in relation to formal employment; second, microfinance is quite developed both quantitatively and qualitatively, reflecting in part what has been described as “the best business environment for microfinance in Latin America” (The Economist, 2009), based on a combination of regulatory framework, investment climate and institutional development. Most of the Peruvian labour force works in microenterprises, either as independent workers (almost half), owner-employers (around one third) or dependent workers (unpaid or paid). As in other Latin American countries, the microenterprise sector is quite heterogeneous by sector, technology and size; many of these differences underlie the also wide heterogeneity in productivity, output, value of sales and labour incomes, as shown by Rodriguez and Higa (2010) for 2004–2008. Over the years the more successful microenterprises have competed and shared markets with other small and medium-sized firms, while others have disappeared. Integrated and specialized NGOs began to provide technical assistance and microcredit to microenterprises during the 1980s and achieved high recovery rates on loans. These experiences were duly noted, and from the 1990s on, taking advantage of more professionals on their staff, governments adopted a more active role towards microenterprises, with policies oriented mostly to providing technical assistance and financial services, mainly through a commercial rather than an assistential approach that included among other things attempts to broaden microenterprises’ access to markets through better information channels and a programme of government purchasing from microenterprises (see below). The focus has been on policies and public–private agreements to provide technical assistance to microenterprises under jointly financed schemes. Many of these public policies have not been large enough to have an aggregate impact; in some cases it remains unclear whether the design and operative mechanisms have made them effective in assisting microenterprises in need of their services. The experience with microfinance has been different. The 1980s saw new, unconventional ways of funding microenterprises, including NGOs, credit and savings cooperatives, municipal funds (Cajas Municipales del Peru—CMACs) created in an effort to decentralize access to financial services and for about a decade the beneficiaries of German technical assistance, which came to compete in different ways with existing informal sources (relatives, suppliers, rotating credit

132

Social protection, growth and employment

and savings associations — ROSCAs). With the economic stabilization policies of the 1990s, the Peruvian development banks, credit and savings cooperatives and other financial institutions collapsed, as did many NGOs, while others were transformed into a new regulated form, Entidad de desarrollo de la pequeña y microempresa (Edpyme). After the collapse of the Banco Agrario, the government launched the Cajas Rurales with private participation in its ownership. In the mid-1990s new regulatory rules for microfinance institutions (MFIs) transformed the whole system, created a new context for the regulated MFIs (CMACs, Cajas Rurales and Edpymes), which contributed to an impressive growth of the loan portfolio, recovery and outreach. The non-regulated MFIs (such as NGOs) and the informal channels were left outside the regulated financial system. During the past decade the profitability of microfinance activity has attracted the attention of commercial banks, which have purchased successful MFIs to transform them into Financieras, which then supply additional financial services. Currently, regulated MFIs co-exist with unregulated ones; a key difference is that the former are legally allowed to take deposits and are supervised by the regulatory institution Superintendencia de Banca y Seguros (SBS) to assure that they fulfill the legal requirements for operation. Fortunately, the SBS has attracted highly qualified and well-paid professionals in charge of implementing and reinforcing the prudential financial regulation. During the past two decades the public sector has also channelled financial resources through its previously reformed second-tier development bank, COFIDE. Today the very modern, regulated microfinance system is expanding rapidly throughout the country; it has participated in mergers and product innovations, designed in part to attract new groups of clients. COFIDE has launched several funding programmes for regulated MFIs. Most recently the public Banco de la Nacion has become involved in microfinance through activities such as its ‘window’ (Ventanilla MyPE) designed to assist regulated MFIs in reaching remote borrowers (see below). Social policies using microcredit as a tool have been practically absent; by law, only regulated financial institutions may deliver even public financial resources. An illustration is the case of Foncodes, which, being a social policy institution, issues microcredit to its target population through the public bank Banco de la Nacion. The MFI sector has thus become quite heterogeneous, with a financially strong regulated MFI in a period of rapid growth (including Cajas Municipales, Cajas Rurales, Financieras, commercial banks and cooperatives) and a search for new client groups beside microenterprises. In addition to this part of the market, there is the non-regulated and heterogeneous sector of NGOs with slow average credit growth, since they cannot mobilize deposits. Other more traditional or informal mechanisms of financing (e.g. money lenders, relatives, suppliers) have persisted historically, supplying finance to the smallest microenterprises and small consumers in both rural and urban areas. Despite the lack of information on this part of the market, it is evident that informal microfinance persists despite all the innovations and regulatory changes and that the continuing heterogeneity among Peruvian microenterprises is matched by differential access to financial services.

Peru

133

Public policy and microfinance Government purchasing from micro- and small enterprises23 In 2003 the government created the Compras del Estado a las MyPEs programme to promote the formalization of micro- and small enterprises (Micro y Pequeñas Empresas—MyPEs) through direct purchases of such goods as school shoes and uniforms (delivered through social programmes such as Foncodes), office supplies, and food for school breakfasts. By law, all public offices of central, regional and local governments must purchase at least 40 percent of their total materials and services from MyPEs, whenever quality is guaranteed and agreed conditions met. Currently MyPEs are among the suppliers of practically all of the items they can offer, including those goods mentioned above and also personal services (e.g. packaging and storage, mail delivery, transportation, catering, tourism) and even some infrastructure projects (e.g. small communications projects). Between 2003 and 2006, this programme remained small, and the legal minimum threshold of 40 percent was not met. Since then it has expanded, and by mid-2009, during the external financial crisis, it had become a centrepiece of the Garcia administration’s policy response. As of mid-2011, official statistics reported more than PEN10 million (almost US$4 million) in purchases from MyPEs, compared to just PEN2 million in 2006. This increase raised the share of government materials and services purchased from MyPEs to 45 percent. The growth coincided with an institutional change, as the programme was moved from the Ministry of Labour to the Ministry of Production. This continues to be a promising strategy to raise the business and labour income of MyPEs. Banco de la Nacion window for microenterprises Given the sort of modern financial system developed over the last two decades in Peru, it might be expected that most financial services for small consumers and microenterprises would be delivered by the (regulated) financial intermediaries, including MFIs. In fact, Peru remains notorious for the absence or near absence of commercial banks and even regulated MFIs in the many remote areas of the country, including most of the Sierra and many small urban centres. In contrast, the government bank, Banco de la Nacion, has offices practically everywhere, including remote locations. This set of conditions gave rise in 2006 to an institutional arrangement between the Banco de la Nacion and any regulated MFI, whereby the latter may use the bank’s premises to issue loans to microenterprises. In that same year another legal change permitted the Banco de la Nacion to act as a second-tier institution providing lines of credit to regulated microfinance institutions, thereby increasing the total supply of finance to MyPEs. The ‘window’ mechanism has expanded throughout the country, and it is generally believed that all the parties benefit because of the expansion of their respective activities. By mid-2011 it was reported that more than PEN1.670 billion (around US$600 million or less than 2 percent of all the microloans) were issued in 500,000 loans under this scheme in the regulated system. Some concerns, not overly plausible, have been voiced by regulated financial institutions afraid of ‘unfair’ competition from a government-owned bank operating in far-flung places seen as market niches.

134

Social protection, growth and employment

The Peruvian microfinance clientèle As indicated above, the coverage and depth of microfinance is relatively high in Peru, with sustained growth of its lending and deposits. By 2011 (MIX Market, 2012), the outstanding portfolio became US$8.8 billion, with 3.6 million borrowers (around 25 percent of the Peruvian population between 15 and 64 years old); the amount of deposits reached US$6.6 billion, with 3.5 million depositors. Among the 70 regulated and formal MFIs, the repayment rates have been kept over 90 percent on average, although some delinquency problems appeared in the smaller MFIs. Deposits have been the primary source of MFI funding, at over 60 percent of the total portfolio. Average loans are around US$2500, and interest rates have fallen in recent years. The Peruvian MFIs have shifted their portfolio allocation somewhat from consumer loans to microenterprise lending, diversifying their basket of financial products (e.g. in 2011 the Caja Municipal Arequipa issued around 40 percent of its loan portfolio to small firms, 25 percent to microenterprises and 20 percent to consumption, according to the Superintendencia de Banca y Seguros).24 Given its active history with microfinance, Peru has naturally been the scene of much discussion and some analysis of the phenomenon. An important early study by Dunn and Arbuckle (2001)25 concluded that “clients can consistently improve enterprise performance and increase household welfare. The positive impacts of microcredit appear to extend to the poor as well, although poor clients were more likely to turn to asset-reducing coping strategies than their non-client counterparts. The results of a separate analysis indicated that for several key variables—including enterprise revenue, enterprise fixed assets, business premise ownership, business licensing, intergenerational launching, and household income—increased time in the microcredit programme was associated with better outcomes on these variables.” They also note that their findings refute a common industry assumption that “borrowers engage in longterm relationships with microfinance organizations and seek ever-increasing loan amounts over time. Only about 40 percent of the original client sample had a current loan in 1999, and they adjusted their loan requests in response to changing circumstances.” Their overall assessment is, “A microfinance industry that provides a full range of services for small transactions may increase the financial options enough to provide the critical boost that low-income households need to achieve their goals.” In her later study, Tedeschi (2008), using a panel data set to test for the impact of microcredit on microenterprise profits while controlling for some of the more worrisome traditional biases (including selectivity in the lending programme) finds that, although such correction does lower the estimated benefits (in line with various earlier studies), it does not erase them; profits are still significantly higher as a result of the loans. In short, the main positive conclusions of the general microfinance literature based on microeconomic evidence26 have been reported for Peru. In a complementary analysis directed at the continuing challenge of monitoring the social effects of microfinance, Copestake et al. (2005) consider in-depth interviews important to gain an adequate understanding of those impacts, and state that most studies of this sort have tended to confirm significant benefits to borrowers.

Peru

135

Among the limitations of micro-level analyses is the difficulty of extrapolating up from it to potential aggregate effects on growth, inequality and poverty. The prevalence of microfinance in Peru and the favourable conditions for its success make this country a natural candidate for analyses of its aggregate impacts on poverty, inequality and growth, as well as economic security. Such analysis is effectively precluded in countries where microfinance is not large enough to make it reasonable to expect its overall or aggregate effects to be identifiable in the context of the various other determinants of these variables. Peru has not only a breadth of microfinance and conditions propitious for its benefits to emerge but an incipient analysis of the dynamics of the informal sector. Two of the possible approaches to increase our capacity to judge actual or potential aggregate effects are (i) econometric analysis in which such outcome variables are regressed against other known or presumed determinants of those variables and also variables that reflect the size and other features of microfinance; and (ii) analysis of the dynamics of the informal sector (and perhaps the adjacent small-firm sector) with a view to better understanding how much productivity growth is taking place and how it occurs, how much it feeds into income increases for those involved, and how microcredit and other variables are involved in this process.

Box 3.2: The new Humala government and proposals for microenterprises and MFIs — a quick glance After the government declared that the entrepreneurial development of MyPEs is crucial for Peru’s economic development, the first policies announced in this area include: • emphasizing the Comprare al Peru programme to increase government purchases from microenterprises; • emphasizing the Marca Peru programme with active participation of microenterprises; and • promoting financial services that may or may not imply an active role by the government in providing financial services:

o promoting financial services through the state-owned or -related banks (Agrobanco, Banco de la Nacion and COFIDE); o emphasizing the use of Banco de la Nacion as a window for other regulated private institutions providing microfinance services; and o explicit lines of credit and a diversified basket of microfinance products. Some concerns have been voiced by larger institutions in operation. How all these policies will be compatible with the economic measures to be implemented is one of the current issues.

136

Social protection, growth and employment

Conclusion Most of the social policies implemented over the last two decades were originally designed to compensate for what were expected to be short-term negative effects on poor people from the economic crisis and the ensuing stabilization policies. In fact, since poverty has been a persistent problem, those policies have ended up as part of the governmentâ&#x20AC;&#x2122;s response to this longer-term problem, complemented by temporary, assistance-oriented tools and approaches. To the extent that poverty is linked to such other long-term problems as underemployment, social policies have increasingly been seen as a way to deal with those problems and to improve the employment possibilities of beneficiaries. There is considerable evidence of Peruâ&#x20AC;&#x2122;s success in reducing poverty at moderate fiscal cost, notwithstanding evidence of leakage of benefits to higher income groups and under-coverage of families within the target group. The SIS is of particular interest and importance because it taps into an enormous social need, reflected in a very rapid increase in coverage after its launch but also in an overburdening of the existing health apparatus, with attendant dangers (quality problems, disaffection of both the public and the health personnel) and associated challenges on the fiscal, administrative and professional staffing fronts. What happens over the next decade will be extremely important, since this programme has very great potential to benefit low-income families, but it will require strong administration and a large spending outlay to achieve that potential. Its challenges are similar to those currently being faced in a number of other developing countries. Another key programme is Juntos, the CCT programme designed to encourage families to keep children in school and to attend health services more frequently. Ex post evaluations have judged it to be successful in lowering poverty and raising educational attendance, without discouraging the work effort of the families involved. On the other hand, there is as yet no evidence of positive impacts on final outcome indicators such as malnutrition. The expected long-term benefits of higher educational outcomes among the children involved will depend on the evolution of labour demand in the economy, which in turn will depend on other aspects of economic policy. Peru now has a large microfinance system, and there is considerable microeconomic evidence of benefits, as well as some aggregate data consistent with there being a significantly positive overall impact on the productivity of microenterprises. More research is needed, however, to confirm such effects and to measure them with greater precision. Some social programmes (e.g. Provias, microfinance) have the potential to generate considerable employment, whereas others do not. We may expect that: â&#x20AC;˘ the programmes that provide items of consumption (e.g. food transfers such as Glass of Milk) or final services (e.g. the literacy programme and SIS) will have very limited capacity to generate employment and labour income in the short term. Whether some of these policies may instead have negative effects by providing incentives for firms

Peru

137

to become informal to hire labour at lower rates is an empirical matter on which little light has been thrown thus far; • CCT programmes can have a large potential to generate better future employment and labour incomes for their beneficiaries, whenever future labour demand is strong enough to make the investment in their education (or health) pay off; • programmes with more direct labour effects are those aimed at creating employment (e.g. ATU, Construyendo Peru) and focusing on repairing roads (e.g. Provias) or smallscale local productive projects (e.g. Foncodes). In terms of employment generation, the longer they run, the larger the effects. All these are by nature temporary programmes, thus the employment they generate is just temporary, usually with a previous training course. The expectation is that after the beneficiaries participate in the programme they will then be hired in the modern sector. This is probably unrealistic in most cases, but if the beneficiary moves on to any decent job, the programme can be considered successful; • in Peru, despite its accelerated growth for almost two decades, employment has grown slowly, while underemployment has been higher that 50 percent. It may explain the persistence of social policies implemented with minor changes by all governments since 1990. This requires a comparative analysis of temporary as well as permanent employment generated by economic activities and by social policies, adapting with methodological approach to the available information; • to better understand both the informal or microenterprise sector itself and the impacts of microfinance, it would be valuable to pursue the analysis into a sufficiently greater depth as to clarify the role of those microfinancial services; and • regarding labour policy per se, Peru undertook a major reform in the early 1990s to simplify labour legislation, reduce the various costs for firms and flexibilize the labour force. There is thus far little evidence of the expected benefits in terms of increased formal-sector employment or other labour outcomes. The possibility remains that effects, both positive and negative, may arise with a substantial lag. This issue requires additional in-depth research.

138

Social protection, growth and employment

Annex 3.1: Economic policies implemented between 1990 and 2010: main features 1990–2000 In 1990, in the middle of changes occurring worldwide (e.g. the fall of the Berlin Wall, the collapse of the Soviet Union, acceptance of the Washington Consensus), the new Fujimori government set the recovery and growth of the Peruvian economy and society as its main goals. The President demonstrated a personal style of government with few advisers and supported by the army. The main features of this period included: • immediate economic policies: elimination of price controls and subsidies, exchange rate controls and interest rate controls, along with macroeconomic adjustment with fiscal discipline and related monetary policies, free trade and capital policies, repayment of external debt; • main economic policy approach: generalized privatization, free trade and foreign capital entrance; • key reforms included the privatization and reduction of government participation in production activities, with mainly foreign firms in the most dynamic activities: mining (gold), commerce, finance (banks) and public services (communications, electricity); • parallel institutional reforms included tax collection, public services regulation and property rights, involving highly qualified professionals back from abroad (Contreras and Cueto, 2004); • fiscal expenditure mostly for physical infrastructure investment and social policy programmes; • labour and pension reforms were oriented to lower labour costs, towards free labour market relations; • financial reforms, including a new banking law; • important consequences to highlight: ○○ Banking institutions and services became more diversified, involving some banks in a downscaling process to reach small and microenterprises as clients, increasing competition to existing Cajas Municipals and other regulated small institutions. ○○ Petrol stations, supermarkets and other sectors became more modern and changed the face of the cities, mainly Lima, run with important foreign capital investment. ○○ Internal demand for residential telephone lines, electricity and gas was largely met, reaching even urban poor families. ○○ There were large flows of dollars, with an implied quick increase in imports (in final goods and inputs), with import businesses flourishing quickly. ○○ Peruvian exports — mostly from mining — grew at high rates, also favoured by external demand.

Peru

139

○○ Other non-traditional labour-intensive agricultural exports expanded quickly due to the spectacular growth of international demand. ○○ Peruvian industry and small agricultural goods were competing against cheaper imports. ○○ Social programmes covered rural and urban areas, with political bias management. By late 1999, the international crisis caused internal Peruvian economic outcomes to deteriorate. After a generalized corruption system linked to the government was discovered, and in the middle of economic crisis and social discomfort, Fujimori resigned (and was later imprisoned).

2000–2001 A transitional government led by Paniagua aimed at undoing the previous corrupt system with new special organisms and implemented institutional and economic reforms to facilitate new elections.

2001–2006 The new President Toledo faced an economy in recession, with low private investment and high expectations from the population, setting political and anti-corruption goals. The main features of this period included: • the general economic policy approach was kept, with: ○○ macroeconomic real and monetary policies oriented to free market economic activities; and ○○ stressing modernization of public policy; • concerns emerged about the growth of financial as well as employment and income outcomes, associated with market policies (Jimenez, 2010). Voices in favour of marketoriented policies argued that no growth was possible in the long term if these policies were stopped (Loayza, 2008); • proposed reforms and medium- to long-term policies could not be implemented because of opposition from society and local populations, mainly against the mining sector. The main policies were: ○○ expansion of privatization programmes in all economic sectors; and ○○ political decentralization reform among the country’s regions; • other policies stimulated private investment in almost any economic sector. The increase in urban residential buildings, modern malls and shopping centres was notable; and • crisis and further recovery of the international economy influenced Peruvian GDP, reinforcing the historical volatility (Schuldt, 2005). The rate of economic growth remained high. Toledo ended his period with very low political acceptance, despite high economic growth.

140

Social protection, growth and employment

2006–2011 Garcia’s second term (after his first one in 1985–1990) stressed new free-market-oriented economic policies, emphasizing the difference and maturity in respect to his past ‘more peopleoriented’ government. The stability of the international economic context until 2008 favoured economic growth, which was not even affected by the 2008–09 world financial crisis. The main features of this period, despite the still limited information available, were: • macroeconomic monetary and real policies reinforced their previous orientation; • private foreign participation in economic activity was stimulated through fiscal incentives; • generalized trade agreements with specific countries were signed with strategic objectives, while general free-trade-oriented policies were reinforced; • accelerated public investment in physical infrastructure and public services in early 2011; • institutional reforms for public goods, property rights and labour reinforced the free market perspective; and • an emphasis on the public management of policies, from design to evaluation, towards government modernization through new approaches (‘budget by results’) and tools (e.g. improved multi-year macroeconomic framework). By July 2011, Garcia lost social and political support, mostly from the poor population, and finished his mandate with accusations of corruption against his party leaders and collaborators.

Peru

141

Annex 3.2: Main social programmes implemented in the 1990s — Fujimori 1990–1992 The first emergency programme initiated by Fujimori was expected to be temporary and was implemented after the structural adjustment policies to mitigate their negative effects on vulnerable groups. The main social programmes included: • Programa de Emergencia Social (PES): social emergency programme; • Foncodes: national fund for social compensation and development, to finance small, local social projects; • Programa Nacional de Asistencia Alimentaria (PRONAA); • Comedores populares: People’s kitchens run by women in poor areas; • Vaso de Leche: Glass of Milk programme to deliver breakfast for children in poor areas; • Cooperacion Popular (COOPOP): reformulated based on the previous government’s design, to implement local programmes; and • Instituto Nacional de Desarrollo (INADES): public office in charge of executing infrastructure projects in poor Peruvian regions.

1993–1995 Social policy was reoriented after the 1993 Paris Club meeting, with Fujimori engaging the Peruvian government to target extreme poverty through focused programmes of education, health, food and employment generation. In addition to the programmes started previously, the main activities and programmes between 1993 and 1995 included: • a national strategy of poverty alleviation is presented (at the Paris Club meeting), setting the basis for the new social policy, to then define new basic guidelines and actions; • the main criterion for social expenditure on infrastructure and the supply of basic goods would be to focus on just the most vulnerable populations; • programmes to promote temporary employment, health, education, security, public services and infrastructure just target extremely poor populations, to improve basic social expenditure and equality in opportunities; • increased social expenditure in education, with a reduced share of total social expenses, in favour of health and other social programmes; • education-related programmes included: ○○ buildings and equipment for public schools across the country (through the National Institute of Educational and Health Infrastructure—INFES);

142

Social protection, growth and employment ○○ educational reforms, with programmes to deal with specific problems in public education (e.g. centralized, with large population coverage but large desertion rates, large share of payroll to total educational expenses); and ○○ transfer of management and funding of public schools to local entities of municipality, parent associations, religious or professional associations, based on the view that they know better the local interests and needs and may also complement the funding; • increased social expenditure in health, mostly to public hospitals with universal coverage, mainly to cover current (payroll) expenses, with minimum investment expenses; and • health-related programmes included: ○○ specific programmes to deal with high infant mortality, high illness rates and undernutrition, previously diagnosed as the main health problems in Peru; ○○ basic health for everybody, reaching the largest coverage among the health programmes; and ○○ transfer of the management of local public health posts to the people, through newly created local committees for shared management (CLAS).

1995–2000 • A narrowed strategy is set, to focus on just the extremely poor population unable to afford a basic food basket. • A poverty-by-district map was built up based on data from several censuses and surveys, emphasizing information about height and weight, to target strategic programmes on those districts. • Strategic programmes focusing on extremely poor districts included: ○○ social assistance programmes: school-age child nutrition, family planning etc.; ○○ social infrastructure programmes: public health, drinking water and drainage systems, basic educational infrastructure etc.; and ○○ economic infrastructure programmes: roads to improve these districts’ access to larger markets, small irrigation and electricity systems etc. • Among the social programmes in education, one highlighted by the government was the programme to improve the quality of Peruvian education (MECEP), which focused on issues about the teaching and learning processes, performance measurement and the modernization of physical infrastructure. Results showed poor outcomes, highly biased to urban, non-poor beneficiaries. • In health, the public social insurance system was reformed under a new institution (ESSALUD) to provide public health services not only to current formal-sector workers but also to their relatives as well as retired people and other population groups.

Peru

143

Annex 3.3: Main social programmes by type — Peru: 2001–2010 2001–2006: Toledoa

2006–2010: Garciab

i. Social programmes: Food and nutrition:

Infrastructure:

- Glass of Milk for defined target population

- the Crecer strategy to fight endemic under-nutrition among children

- People’s Dining

- Glass of Milk redefined

- Pronaa (various programmes for food assistance, e.g. school breakfast, children dinning)

- People’s Dining redefined

- Rural electrification

- Local population groups implementing projects (= redefined nucleos ejecutores—Foncodes)

- Rural Provias for improving roads - Pronamachs (for hydric basin management and soil preservation activities) - Various urban residential programmes (e.g. Techo Propio) - Improve your life and other Foncodes projects for social and economic infrastructure in rural areas Temporary work:

- Pronaa reformulated - conditioned nutrition programmes (under the Juntos programme)

- Decentralized Rural Provias - Pronamachs - Various urban residential programmes (with Banco de Materiales) - projects for social and economic projects in rural areas—Foncodes

- Rural: Let’s work (A trabajar rural)

- Rural: Let’s produce (A producir rural)

- Youth job training (Pro-joven)

- Urban: Let’s work (A trabajar urbano)

- Female job insertion (Profece)

- Mi Empresa (merging the two PromPyME and Profece programmes)

- Programmes for supporting microenterprises (Promocion de la pequeña y la microempesa— PromPyME)

- Technical assistance and information services for microenterpises (CreceMyPE)

ii. Education: - National literacy programme - Improving elementary education - Improving initial and secondary education - Rural education programme - Huascaran programme

- National literacy programme reinforced - Improved children’s school attendance (under the Juntos programme) - Permanent training and evaluation of public school teachers

Source: author’s elaboration, based on information from other sources, as indicated. a Information taken mostly from Vasquez (2006: 14, Figure 1.1; and 52–55, Annex) and Du Bois (2006: 22–23, Figure 2.1). b Information taken from various sources including Portocarrero et al. (2010: 275–325).

144

Social protection, growth and employment 2001–2006: Toledoa

2006–2010: Garciab

iii. Health: - Support for improving the health sector - Improved health assistance (under the (PAR Salud) Juntos programme) - Integrated Health Insurance (Seguro Integral de Salud— SIS)

- Reformulated Integrated Health Insurance (SIS redefinido)

iv. Other important experiences: Multidimensional programmes, involving several public sectors and offices

- Rural ownership legalization - Conadis (for disabled people) - National child care for at-risk children (Wawa-wasi) - Juntos, a cash transfer programme started in 2005

- Reformulated social policy approach, with changes in poverty reduction strategies (e.g. Plan de reforma de los programas sociales) - Defined political authorities in charge of coordinated social policies (e.g. CIAS) - Juntos strengthened - Mandatory use of official identification document (DNI) for adulthood - Enforced issue and use of free identification document (DNI) for children

Peru

Annex 3.4: Explaining household participation in a social programme — a Logit regression marginal effects Participate in a social programme

2009

2005

2000

0.07735*

0.06815*

0.06071*

36.95

33.92

10.68

0.13626*

0.12698*

0.06424*

17.74

15.61

3.03

-0.14454*

-0.19620*

-0.13993*

-18.48

-24.89

-6.89

0.01054

0.03317*

0.02606

1.4

4.05

1.33

.-0.00305*

.-0.00251*

.-0.00241*

-11.71

-9.03

-3.42

.-0.27389*

.-0.32688*

.-0.23815*

-29.04

-32.21

-8.95

.-0.03891*

.-0.06280*

-0.00869

-9.26

-12.51

-0.76

22985

21287

2929

LR chi-squared (7)

5639.77

5935.96

397.82

Prob > chi-squared

Pseudo R-squared

0.1939

0.2135

0.1062

-11724.609

-10933.649

-1673.4231

Number of dependents Poverty status

Urban area2 Works in a microenterprise3 Age of head of household Male head of household

Education level

# Observations

Log likelihood

Notes: 1 1=Poor or extremely poor, 0= non-poor 2 1=Urban, 0=Rural 3 1=1 to 10 workers, 0= 11 and more 4 1=Male, 0=Female 5 Includes both complete and incomplete level of education * Significant at 0.1 Source: author’s elaboration, based on the ENAHO for 2000, 2005 and 2009.

145

146

Social protection, growth and employment

Notes 1 Mineral dependency tends to discourage agricultural and manufacturing activities (the other main tradables) and thus leads to a focus on services. 2 See Chapter 1 for definitions of these terms as used in this study. 3 For the logic of such reforms, see the chapters in Heckman and Pagés (2004). 4 Between the previous peak (1987) and 2010, growth averaged just 3.0 percent, highlighting the importance of the fast growth after the sharp downturn of 1987–1992. GNI per capita grew at 1.3 percent per year. 5 This view is broadly consistent with that which emerges from the most in-depth study thus far of informality in Latin America, that of Perry et al. (2007). 6 Although there are no comparable income distribution or poverty data from the early 1980s, per capita income was higher then than at the beginning of the past decade; the absence of any reason to believe that inequality had fallen significantly over that interlude suggests that poverty was probably the same or higher than two decades before. 7 Comparing these figures to those of other countries of the Latin America and Caribbean (LAC) region as well as to economies that share similar socio-economic characteristics, it is notable that Peru’s social expenditure is higher than that of Mexico, the other country of the region addressed in this publication, which only spends 7 percent of GDP on social expenditure. Nevertheless, Peru’s social expenditure is still very far away from public social expenditure invested by Organisation for Economic Co-operation and Development (OECD) countries, which amounted to 21 percent of GDP on average in 2005, with Sweden and France even spending 29 percent of GDP. Taking the important role of social expenditure in the context of LAC into account, aiming at bridging the wide disparities in living standards between rich and poor people in the region with the highest social inequality worldwide, the minimal growth of social expenditure relative to GDP in Peru within the given timeframe does not appear adequate to address the country’s giant income gap. One explanation for this might be the high public debt ratios, which remain at over 50 percent of GDP in the entire region, limiting the ability of governments to meet social needs. 8 Previsional expenditures are social expenditures to provide basic social services to older persons. 9 Dividing public social expenditure further into different sectors, the three biggest categories of social transfers as implemented by OECD member states are pensions (on average 7 percent of GDP), health (6 percent) and income transfers to the working-age population (4 percent). Public spending on other social services exceeds 5 percent of GDP only in the Nordic countries, where the public role in providing services to elderly and disabled people and families is the most extensive. Comparing these data to Peru (Table 3.3), most basic social expenditure flows into the sector of health and sanitation (PEN45.7 million), closely followed by education and culture (PEN40.5 million), and only a relatively small amount being invested in social protection and pensions (PEN13.8 million) (data from 2010). 10 The trends may also be biased due to redefinitions of some programmes. 11 Note, however, that by all estimates poverty did fall rapidly from local peaks reached around 2002, albeit from very high levels. A further complication in using poverty figures to evaluate any social programme whose main effect is not to increase the monetary income of families is that its effects cannot be expected to show up in that monetary income, which is the basis of most poverty estimates. Chacaltana (2006: 14–16) questions the large targeting distortions often attributed to several social

Peru

147

programmes, arguing that they are substantially due to measurement problems inherent in the coverage of the data source (ENAHO), such as a too small rate of response that opens the door for nonrepresentativity; to differing implicit or explicit definitions of ‘beneficiary’ between families and the social programme staff; and to a focus on the number of non-poor families that are beneficiaries, rather than on how many resources are directed to them. 12 On this issue see also the World Bank’s evaluation which identifies poor targeting, deficient financial management and funding leakages as factors that have undermined the efficiency of pro-poor public spending (López-Cálix et al., 2002). 13 Indicators such as the number of public complaints and the size of staff involved relative to tasks carried out are consistent with this view. 14 Other programmes may be equally or more important when judged by other criteria. 15 This is often the case. As highlighted, for example, by Shaffer (2003), ways to save time and effort expended in domestic tasks may be the most important benefits from such projects. 16 For the results of the first impact evaluation, see Perova and Vakis (2009). 17 The importance of job creation in countries such as Peru is usually not so much to pull people out of open unemployment as to raise aggregate demand for labour such that total wage earnings can rise. 18 The incidence of the benefits of social spending are estimated on the basis of the household survey (of 2009) which tabulates household responses as to what programmes and services they drew on. This direct information on incidence of benefits probably means that the results reached are more accurate than in some other studies where less direct approaches have to be used. 19 This figure differs somewhat from that implicit in Table 3.4, which is 4.3 percent of GDP when administrative expenses are excluded and about 5.8 percent when those expenditures are included. 20 As with most such estimates of the incidence of government budgets, the effects of items such as infrastructural investment (which are very hard to estimate) are not taken into account, and there is no general equilibrium analysis to trace the indirect effects following from the budget’s first round effects. 21 Government (all levels) revenues as a percentage of GDP are far higher in Brazil (at over 50 percent) and in Bolivia (close to 40 percent) than in Peru and Mexico at just over 20 percent. 22 See, especially, Guerriero (2012) and also Rodriguez and Jayadev (2010). Evidence for Paraguay is presented by Gonzalez et al. (2010: 439). The fact that this pattern has characterized most industrial countries during the period of globalization also suggests that it may also have been occurring in some developing countries. 23 In legal terms, a MyPE is defined as a natural or legal firm under any organizational form, operating with 1 to 10 workers and annual sales value up to 150 tax units (as of 2007 one tax unit was equal to PEN3450 or around US$1200), while small enterprises had 11 to 50 workers and annual sales up to 850 tax units (Mintra, 2006). 24 www.sbs.gob.pe . 25 These authors made an attempt to take into account the methodological limitations to their analysis, including failure to measure displacement effects. 26 For a recent review, see Armendariz and Morduch (2010).

148

Social protection, growth and employment

References Alcazar, L. and E. Wachtenheim, Determinantes del funcionamiento de los proyectos de Foncodes, Instituto Apoyo, Lima, 2001. Apoyo Consultoria, Situacion de las fusiones de los programas sociales: alcances y resultados, Report prepared for GTZ, Apoyo Consultoria, Lima, 2009. Armendáriz, B. and J. Morduch, The Economics of Microfinance, Second Edition, MIT Press, Cambridge, MA, 2010. Arrospide, M., Presupuesto publico evaluado: Programa Nacional de Apoyo Directo a los mas Pobres Juntos, Presidencia del Consejo de Ministros, Lima, 2009. Arrospide, M., K. Rozas and J. Valderrama, Presupuesto evaluado—Seguro Integral de Salud, Ministerio de Economía y Finanzas, Lima, 2009. Berry, A., Putting economic growth at the service of human development, draft, 2010a. Berry, A., Concept note, draft, 2010b. Berry, A., Growth, employment and distribution impacts of mineral dependency: four case studies, draft, 2010c. Berry, A., Por qué los países no tienen verdaderas estrategias de empleo e ingresos laborales?, draft, 2010d. Bonoli, G., ‘Classifying welfare states: a two-dimension approach’, Journal of Social Policy, 2000, 26(3): 351–372. Chacaltana, J., Impacto del programa A Trabajar Urbano: ganancias de ingresos y utilidades de las obras. Informe final, CEDEP, Lima, 2003. Chacaltana, J., ¿Se puede prevenir la pobreza? Informe final de investigación, Centro de Investigacion Economica y Social (CIES), Lima, 2006. Chong, A., J. Galdo and J. Saavedra, ‘Informality and productivity in the labor market in Peru’, Journal of Economic Policy Reform, 2008, 11(4): 229–245. Contreras, C. and M. Cueto, Historia del Perú Contemporáneo. 3rd Edition, PUCP, UP and IEP, Lima, 2004. Controlaría General de la Republica, Programas sociales en el Perú: elementos para una propuesta desde el control gubernamental, Contraloría General de la Republica and Cooperación Alemana al Desarrollo, Lima, 2008.

Peru

149

Controlaría General de la Republica, Departamento de Programas Sociales, Informe Macro sobre la Gestión del Programa del Vaso de Leche, Informe No. 724-2011-CG/Psc-IM, Controlaría General de la Republica, Lima, 2011. Copestake, J., P. Dawson, J.P. Fanning, A. McKay and K. Wright-Revolledo, ‘Monitoring the Diversity of the Poverty Outreach and Impact of Microfinance: A Comparison of Methods Using Data from Peru’, Development Policy Review, 2005, 23(6): 703–723. Cotlear, D. (ed.) Un nuevo contrato social para el Perú: ¿como lograr un país mas educado, saludable y solidario?, World Bank, Lima, 2006. Díaz, J., ‘Estructura de ingresos en Lima Metropolitana: 1986–1995’, CIES Investigaciones Breves, No. 11, Centro de Investigacion Economica y Social (CIES) and GRADE, Lima, 1999. Díaz, J., Employment and Earnings Trends in Peru, 1997–2008: Evidence from Household Surveys, GRADE, Lima, 2009. Dornbusch, R. and S. Edwards (eds), The Macroeconomics of Populism in Latin America, University of Chicago Press, Chicago, IL, 1991. Du Bois, F., Programas sociales, salud y educación en el Perú: un balance de las políticas sociales — democratizando el gasto, IPESM and Konrad Adenauer Stiftung, Lima, 2006. The Economist (2009) ‘Ripe for development: Peru takes the top slot in a ranking of microfinance business environments in Latin America and the Caribbean’, The Economist website, 19 Feb. 2009, www.economist.com/node/13097723 (accessed 12 Dec. 2012). Escobal, J., ‘The Role of Public Infrastructure in Market Development in Rural Peru’, Thesis, Wageningen University, Wageningen, 2005. Escobal, J. and C. Roman, ‘Enhancing Income Opportunities for the Rural Poor: The Benefits of Rural Roads’ in J.M. Fanelli and L. Squire (eds), Economic Reform in Developing Countries: Reach, Range, Reason, Edward Elgar, Global Development Network Series, Cheltenham, UK, 2008. Escobal, J. and C. Ponce, ‘Desigualdad espacial en el Peru en las ultimas tres decadas’, Conference ‘30 años de aniversario de GRADE’, GRADE, Lima, 8 Nov. 2010, www.grade.org.pe/download/Presentacion_Javier_Escobal.pdf (accessed 12 Dec. 2012). Escobal, J., Saavedra, J. and M. Torero, ‘Los activos de los pobres en el Peru’, Documento de trabajo, No. 26, GRADE, Lima, 1998. Figueroa, A., La Sociedad Sigma: una teoría del desarrollo económico, Fondo Editorial — Pontificia Universidad Católica del Perú and Fondo de Cultura Economica, Lima, 2003.

150

Social protection, growth and employment

Figueroa, A., ‘El problema del empleo en una sociedad Sigma’, Working Paper, No. 249, Pontificia Universidad Católica del Perú, Department of Economics, Lima, 2006. Fondo de Cooperacion para el Desarrollo Social—Foncodes, ‘Reseña Historica’, Lima, 2012. www.foncodes.gob.pe/portal/index.php/nosotros/nosotros-resena (accessed 4 Jan. 2013). Francke, P., ‘Politicas sociales: balance y propuestas’, Working Paper, No. 194, Pontificia Universidad Católica del Perú, Department of Economics, Lima, 2001. Francke, P and J. Espino, ‘Los programas de inversión social y la generación de empleo: el caso de Foncodes’, Diagnostico y Propuesta, No. 11, Centro de Investigacion Economica y Social (CIES), Lima, 2001. Gana Peru, Lineamientos centrales de política económica y social para un gobierno de concertación nacional, campana Ollanta Presidente, 2da. Vuelta, unpublished version, 2011. Giugale, M., V. Fretes and J. Newman, Peru: la oportunidad de un pais diferente: prospero, equitativo y gobernable, World Bank, Lima, 2006. González, C., J. Ramírez and M. Perera, ‘El Mercado Laboral Paraguayo: Un Análisis del Comportamiento del Empleo y los Ingresos entre 1997 y 2008’ in J. Rodríguez and A. Berry (eds), Desafíos Laborales En América Latina Después De Dos Décadas De Reformas Estructurales: Bolivia, Paraguay y Perú, 1997–2008, Instituto de Estudios Peruanos, International Development Research Centre and Pontificia Universidad Católica del Perú, Lima, 2010: 425–522. Gonzales, E., ‘¿Esta cambiando el Peru? Crecimiento, desigualdad y pobreza’, Oxfam: Pobreza, desigualdad y desarrollo en el Peru—Informe Anual 2007–2008, Oxfam, Lima, 2008: 54–61. GRADE, Investigacion, política y desarrollo en el Peru, GRADE, Lima, 2007. GRADE, Consultoria para el análisis de la evidencia acerca del Programa Provias Descentralizado, periodo 2000–2009, GTZ, Governance, Democracy and Inclusion Programme, Lima, 2010. Guerriero, M., ‘The Labour Share of Income around the World: Evidence from a Panel Dataset’, Development Economics and Public Policy Working Paper Series, No. 32/2012.r, University of Manchester, Institute for Development Policy and Management (IDPM), Manchester, UK, 2012. Heckman, J. and C. Pagés, Law and Employment: Lessons from Latin America and the Caribbean, University of Chicago Press, Chicago, IL, 2004. Iguiñiz, J., ‘Lucha ¿contra que pobreza?’, Revista Economía, 2005, 28(55–56): 159–186.

Peru

151

Instituto Nacional de Estadistica e Informatica (INEI), Informe Tecnico de la Evaluación de la Pobreza, INEI, Lima, 2009. Instituto Trabajo y Familia, Tres años promoviendo el desarrollo integral de la población rural en situación de extrema pobreza, UNDP, Lima, 2009. Inter-American Development Bank (IDB), ¿Los de afuera? Los patrones cambiantes de exclusion en America Latina y el Caribe, IDB, Washington, DC, 2008. Jaramillo, M. and J. Saavedra, ‘La desigualdad en el Peru despues de la reforma structural 1997–2006’, Conference ‘30 años de aniversario de GRADE’, GRADE, Lima, 8 Nov. 2010, www.grade.org.pe/download/Presentacion_Miguel_Jaramillo.pdf (accessed 12 Dec. 2012). Jaramillo, M., Baanante, J. and T. Sanz, Construyendo Peru — Informe Final: Presupuesto evaluado, Ministerio de trabajo y promoción social, Lima, 2009. Jiménez, F., La economía peruana del último medio siglo: ensayos de interpretación, Pontificia Universidad Católica del Perú, Department of Economics and CISEPA, Lima, 2010. Lavado, P., ‘Desigualdad en los programas sociales en el Perú’, Diagnostico y Propuesta, No. 34, Centro de Investigacion Economica y Social (CIES) and World Bank, Lima, 2007. Leon, J., ‘Pobreza y desigualdad distributiva’, Diagnostico y Propuesta, No. 40, CIES, CEDEP, PUCP, Macroconsult and IEP, Lima, 2008. Levy, S., Good intentions, bad outcomes: social policy, informality and economic growth in Mexico, The Brookings Institution, Washington, DC, 2008. Levy, S. and E. Rodriguez, Sin herencia de pobreza. El programa Progresa-Oportunidades de Mexico, Inter-American Development Bank, Mexico City, 2005. Lizarzaburu, P., Apreciacion final del programa A Trabajar Urbano 2002–2006. SSA No. 014/007, United Nations, Lima, 2007. Loayza, N., ‘El crecimiento económico en el Perú’, Revista Economía, 2008, 31(61), January–June. López-Cálix, J.R., L. Alcázar and E. Wachtenheim, Improving the Efficiency of Public Spending in Peru, World Bank, Washington, DC, 2002. Lopez-Calva, L.F. and N. Lustig, Declining Inequality in Latin America: A decade of Progress?, Brookings Institution Press, Washington, DC, and UNDP, New York, 2010. Lustig, N., ‘La reducción de la desigualdad en America Latina: ¿una decada de progreso?’, Conference ‘30 años de aniversario de GRADE’, GRADE, Lima, 8 Nov. 2010, www. grade.org.pe/download/Presentacion_Nora_Lustig.pdf (accessed 12 Dec. 2012).

152

Social protection, growth and employment

Lustig, N., ‘Fiscal policy and income redistribution in Latin America: challenging the conventional wisdom’, Tulane Economics Working Paper Series, No. 1124, Tulane University, New Orleans, LA, 2011. Ministerio de Educacion, Programa Nacional de Movilizacion para la Alfabetizacion, Ministerio de Educacion, Lima, 2009. Ministerio de la Mujer y el Desarrollo Social (Mimdes), Compendio Estadistico, Tabla 2.18: Beneficiarios atendidos a través de proyectos concluidos del Foncodes por línea de acción según departamento, Mimdes, Lima, 2010. Ministerio de Trabajo y Promocion del Empleo (Mintra), La microempresa: una propuesta metodológica y ejercicio de aplicaicon en Lima Sur. BEL 34, Año 9—PEEL, Mintra, Lima, 2006. Ministerio de Transportes y Comunicaciones (MTC), Estudio definitivo de ingeniería para la rehabilitación y mejoramiento de la carretera Patahuasi-Yauri-Sicuani-Tramo San Genaro- El Descanso, km. 11+096 – km. 42+860, Provias Nacional, Lima, 1999. Ministerio de Transportes y Comunicaciones (MTC), Informe de recomendaciones del presupuesto evaluado — Conservacion o mantenimiento de carreteras, Direccion Nacional de Presupuesto Publico, MTC, Lima, 2009. MIX Market, Microfinance in Peru: Country Briefing, MIX Market, Washington, DC, 2012. Monge, A., E. Vásquez and D. Winkelried, ‘¿Es el gasto público en programas sociales regresivo en el Peru?’, Diagnostico y Propuesta, No. 84, Centro de Investigacion Economica y Social (CIES) and Universidad del Pacífico, Lima, 2009. Morales, R., J. Rodríguez, M. Higa and R. Montes, ‘Transiciones laborales, reformas estructurales y vulnerabilidad laboral en el Perú: 1998–2008’ in J. Rodriguez and A. Berry (eds), Desafios laborales en America Latina después de dos decadas de reformas estructurales: Bolivia, Paraguay, Peru (1997–2008), Instituto de Estudios Peruanos, International Development Research Centre and Pontificia Universidad Católica del Perú, Lima, 2010. Moron, E., Identificación de beneficiarios de Programas sociales: opciones y limites del SIsFoh en Peru, draft, Ministerio de Economía y Finanzas, Viceministerio de Economia, Lima, 2009. Páez, A., ‘Peru: Glass of Milk Half Empty’, Inter Press Service News Agency website, 14 May 2010, www.ipsnews.net/news.asp?idnews=51436 (accessed 12 Dec. 2012). Palma, C., Programa de Vaso de Leche en el Peru, Ponencia en la Facultad de Medicina, Seminario de Salud Publica — Universidad Particular San Martin de Porres, Lima, 2008.

Peru

153

Parodi, C., Economía de las políticas sociales, Centro de Investigacion de la Universidad del Pacifico, Lima, 1997. Parodi, C., Peru 1960–2000: Politicas economicas y sociales en entornos cambiantes, Centro de Investigacion de la Universidad del Pacifico, Lima, 2000. Perova, E. and R. Vakis, The longer the better: Duration and Program Impacts of Juntos in Peru. Research Report, World Bank, Washington, DC, 2011. Perry, G.E., O. Arias, H. Lopez, W.F. Maloney and L. Serven, Reduccion de la pobreza y crecimiento: circulos virtuosos y circulos viciosos, World Bank, Washington, DC, 2006. Perry, G.E., W.F. Maloney, O. Arias, P. Fajnzylber, A. Mason and J. Saavedra-Chanduvi, Informality, Exit and Exclusion, World Bank, Washington, DC, 2007. Portocarrero, F., E. Vasquez, and G. Yamada, Politicas sociales en el Peru: nuevos desafios, Centro de Investigacion de la Universidad del Pacifico, Lima, 2010. Riveros, L.A., International differences in wage and nonwage labor costs, World Bank, Washington, DC, 1989. Rodriguez, F. and A. Jayadev, ‘The Declining Labor Share of Income’, Human Development Research Paper, No. 2010-36, UNDP, New York, 2010. Rodriguez, J., Diferencias en los ingresos laborales de los trabajadores asalariados y los autoempleados en el Peru, 2008–2009, draft, Pontificia Universidad Católica del Perú, Lima, 2011. Rodriguez, J. and A. Berry (eds), Desafios laborales en America Latina después de dos decadas de reformas estructurales: Bolivia, Paraguay, Peru (1997–2008), Instituto de Estudios Peruanos, International Development Research Centre and Pontificia Universidad Católica del Perú, Lima, 2010. Rodríguez, J. and M. Higa, ‘Informalidad, empleo y productividad en Perú’, in J. Rodriguez and A. Berry (eds), Desafíos Laborales En América Latina después de dos décadas de reformas estructurales: Bolivia, Paraguay y Perú, 1997–2008, Instituto de Estudios Peruanos, International Development Research Centre and Pontificia Universidad Católica del Perú, Lima, 2010. Saavedra, J. and J. Diaz, ‘Desigualdad del ingreso y del gasto en el Perú antes y después de las reformas estructurales’, Reformas Economicas Series, No. 34, Economic Commission for Latin America and the Caribbean, Santiago de Chile, 1999. Schuldt, J., Bonanza macroeconómica y malestar microeconómico: apuntes para el estudio del caso peruano 1988–2004. 1st Edition, Centro de Investigación de la Universidad del Pacífico, Lima, 2005.

154

Social protection, growth and employment

Seinfield, J. and K. La Serna, ‘Avanzando hacia el aseguramiento universal: como lograr la protección en salud de personas de ingresos medio bajos y bajos?’, Economia y Sociedad, No. 64, Centro de Investigacion Economica y Social (CIES), Lima, 2007. Shaffer, P., ‘Difficulties in Combining Income/Consumption and Participatory Poverty Approaches: Issues and Examples’ in R. Kanbur (ed.), Qualitative and Quantitative Poverty Appraisal, Permanent Black, Delhi, 2003. Sheahan, J., La economía peruana desde 1950: buscando una sociedad mejor, IEP, Lima, 2001. Stifel, D. and H. Alderman, ‘The “Glass of Milk” Subsidy Program and Malnutrition in Peru’, The World Bank Economic Review, 2006, 20: 421–448. Suarez, M., Caracterizacion del Programa del Vaso de Leche — Reporte, Ministerio de Economia, Direccion General de Asuntos Economicos y Sociales, Lima, 2003. Tanaka, M. and C. Trivelli, ‘Las trampas de las focalizacion y la participación: pobreza y politicas sociales en el Peru durante la decada de Fujimori’, Working Paper, No. 121, Instituto de Estudios Peruanos, Lima, 2002. Tedeschi, G.A., ‘Overcoming Selection Bias in Microcredit Impact Assessments: A Case Study in Peru’, Journal of Development Studies, 2008, 44(4): 504–518. Thorp, R., Economic management and economic development in Peru and Colombia, University of Pittsburgh Press, Pittsburgh, PA, 1991. Trivelli, C., ‘Pobreza y políticas sociales’, Economía y Sociedad, No. 70, Centro de Investigacion Economica y Social (CIES), Lima, 2008. Valdivia, M., ‘Contracting the Road to Development: Early Impacts of a Rural Roads Program’, PEP-PMMA Cahiers de Recherche, No. 2010-18, Partnership for Economic PolicyPoverty Monitoring, Measurement and Analysis, Quebec, 2010. Vasquez, E., ‘Programas sociales y pobreza’, Aportes para el gobierno peruano 2006–2011, Centro de Investigacion Economica y Social (CIES) and Universidad del Pacífico, Lima, 2006. Vasquez, E. and D. Winkelried (eds), Buscando el bienestar de los pobres: ¿cuan lejos estamos?, Centro de investigación de la Universidad del Pacifico, Lima, 2003. Vasquez, E., R. Cortez, C. Parodi, J. Montes and G. Riesco, ‘Peru’, Gasto publico en servicios sociales básicos en America Latina y el Caribe: Analisis desde la perspectiva 20/20, CEPAL, Santiago, 1999. Verdera, F., La pobreza en el Peru: un analisis de sus causas y de las politicas para enfrentarla, Fondo Editorial — Pontificia Universidad Católica del Perú, Clacso and Instituto de Estudios Peruanos, Lima, 2007.

Peru

155

Webb, R., ‘Políticas gubernamentales y distribución del ingreso en el Perú: 1961–1971’ in R. Webb and A. Figueroa (eds), Distribución del Ingreso en el Perú, Instituto de Estudios Peruanos, Lima, 1975. World Bank, Poverty reduction and growth: virtuous and vicious circles, World Bank, Washington, DC, 2006a. World Bank, Understanding socioeconomic and political factors to impact policy change, World Bank, Washington, DC, 2006b. World Bank, ‘Peru Rural Road Program: Bringing Access to the Rural Poor, IBRD Results, March 2010. Yokoyama, E., ‘Leakage and Accountability in Service Delivery: Peru’s “Glass of Milk” Child Nutrition Program’, Thesis submitted for the degree of Master of Public Policy, Georgetown University, Washington, DC, 2007.

4 Challenges of transition economies: Economic reforms, emigration and employment in Tajikistan Omair Amir and Albert Berry Background history and recent scenario Tajikistan is a small, low-income country, which over the last 20 years has gone through a traumatic transition following the break-up of the former Soviet Union, an ensuing civil war and the severe economic crash that accompanied these two events. It was left with a ravaged economy and a weak institutional base from which to manage its halting transition toward a more marketoriented economy. The independence that came in 1991 was extremely costly; previously fully integrated into the Soviet Union and the recipieint of considerable net subsidies that helped to prop it up, Tajikistan was left without those benenfits but with the legacy of a civil war that raged on until 1997, when a peace deal between the two parties was brokered with the help of the United Nations and neighbouring countries. Tajikistanâ&#x20AC;&#x2122;s unusual history and setting have powerfully influenced its economy. A very remote region with a mostly mountainous landscape and no access to seaports, it had the lowest income of all the republics of the Soviet Union and has the lowest income of the current Commonwealth of Independent States (CIS) countries. Although some heavy industry was moved to the region during the Second World War, Moscow mainly dedicated the Tajik Soviet Socialist Republic (SSR) to growing cotton. To that end, people were relocated and a vast irrigation network set up to increase yield and output. In the 1980s the Tajik SSR boasted some of the worldâ&#x20AC;&#x2122;s highest-yielding cotton, which was noted also for its very good quality (long staple). Based on

158

Social protection, growth and employment

this marked comparative advantage within the USSR, Tajikistan became so specialized that it grew almost nothing else, food included, thereby setting the stage for later food shortages during the civil war and leaving food insecurity as one of the most pressing challenges facing the country today. Much of the infrastructure that survived the Soviet collapse was devastated during the civil war. While neighbouring republics were on their way to making a reasonably successful transition, Tajikistan’s economy spiralled downwards, with official gross domestic product (GDP) bottoming out in 1996 at nearly 70 percent below its 1991 level. Industrial output, most of which had arisen within a centralized command structure, was largely shut down. Agricultural productivity suffered immensely, as machinery was destroyed and lands abandoned. Food grown from personal plots, together with food aid from donor nations, were the only defences against starvation, as the collectivized agricultural land was almost entirely devoted to growing cotton, and people lacked the income to buy enough imported food. It was not until the early 2000s, with the restoration of peace allowing much-needed macroeonomic stability, that the economy began to show signs of consistent growth, and some underemployed resources were brought back into use. To all intents and purposes, Tajikistan’s economic transition from a centrally planned system toward a more market-oriented system only began around 1997; most of the ensuing period has been as much a story of recovery from war as of that transition. Meanwhile, the world energy crisis, climate change and demographic pressures have all had adverse effects on this landlocked, food-importing nation. The world food crisis of 2008 brought belated recognition of the vulnerabilities of such countries as Tajikistan.1 New food price spikes of 2011 signalled that volatility in food markets is likely to be a permanent feature of the global macro environment for the foreseeable future and that policy needs to reflect this unfolding reality. Analysis of labour markets, social protection and the interface between them must be undertaken against the background of the above facts. Since a majority of Tajikistan’s population (and, particularly, poor people) live in rural areas that grow cotton, societal welfare is inevitably linked to this product. Much of the vulnerability of Tajikistan’s poor people is due to the institutional collapse of the collective farming sector, which has forced the economy to redirect a good share of its basic resource—labour—to employment outside its borders, in particular in Russia. The extent of this switch is reflected in Tajikistan’s very high ratio of remittances to GDP, reaching an estimated 30–50 percent in 2008. Given the massive inability to create remunerative employment at home, Tajikistan has been fortunate to have access to a market in Russia for primarily young male immigrants. Their remittances have become the economic mainstay of many receiving families, as well as being an indirect boon to others, as the circulation of the remittances within the domestic economy encourages local production of services and other non-tradables. On the other hand, the emigration/remittance phenomenon creates new vulnerabilities and concerns for Tajikistan. Remittances are not available to all families and can be an unstable source of income at both family and aggregate levels, due in part to economic factors (witness the decline in remittances in 2009 as a result of the global financial crisis) and in part to political ones.

Tajikistan

159

The social protection challenge involves developing a system that complements and eases the welfare problems characteristic of such a remittance-based system, while also addressing the needs of the families for whom this source of income does not provide support. The severity of the challenge derives from the low average income level; the particular need to defend people against the exigencies of an unfamiliar market economy, including especially food insecurity (a traditional problem in this country); and the fallout from the population movements implicit in such a high level of emigration. As in any low-income country, an important first step is to generate pro-poor growth, which means strong employment-creating growth. Tajikistan has not been successful in this regard, partly due to circumstances exogenous to the country and to its policy makersâ&#x20AC;&#x201D;the severe macroeconomic crash and the loss of capital during the civil war, and the decline of the dominant cotton sector since the days of the Soviet Union. Complementing these factors has been the governmentâ&#x20AC;&#x2122;s failure to expedite a transfer of land to small farmers (former workers) so that they could in the first instance produce food for subsistence and in the second develop a diversified, productive and employment-creating small commercial agriculture. A challenge for the future is that a number of the hoped-for motors of growth and sources of foreign exchange appear to be capital-intensive and hence not job-creating. If this proves to be the case, then even as growth continues and other products overtake cotton as major exports, a severe employment challenge will remain, which will continue or intensify the need for good social protection policies. In short, Tajikistan now faces strong and overlapping challenges in the areas of growth, employment generation/inequality and social policy. Action on each of these objectives must be closely coordinated for success in each to be achieved, especially in the areas of employment and social protection. Maintaining growth of the domestic economy to the point where per capita incomes quickly recover their pre-crisis peak is a major challenge given the currently low investment rate, the unattractive setting for investors, local and international, and the tight monopolistic control of resources exercised by a group of insiders, which deters employment creation and prevents or discourages the entry of competitors in some industries, such as cotton processing and trade. And even if the rate of growth is reasonably adequate, it may not be sufficiently employment-creating for its benefits to trickle down to the population. As noted, the capital intensity of such planned motors of future growth as energy development, minerals etc. implies limited potential for job creation. A more general factor, noted in several other countries of the former Soviet Union, is the strongly monopolistic economic structure that naturally emerges from the privatization of the large production units that comprise a centralized system. State enterprises are replaced by a few large private or pseudo-private monopolists, leading to a concentration of capital income and to monopsonistic behaviour that keeps prices low to producers (such as cotton farmers) and thus lowers the potential for job creation. Under a central planning system, the state takes responsibility for allocating everyone a job somewhere in the economy. In a market system there is no institution to perform that function; as a result, in most developing countries the informal sector absorbs much of the labour that has no decently remunerative job alternatives. Success on the employment front in Tajikistan will depend on small-scale agriculture (cotton, production for home consumption, and other) and on labour-absorbing (usually smaller-scale) activities outside agriculture. In

160

Social protection, growth and employment

former Soviet bloc countries, the job-creating potential of informal activities is less evolved than elsewhere in the developing world for various reasons, including a relative lack of entrepreneurial experience and the only gradual development of the informal markets in which microenterprises can sell. It is somewhat unclear how much future demographic developments will accentuate the employment challenge from the labour supply side, where there is currently a major surplus. Tajikistan was in the midst of a bout of very fast population growth (around 3 percent per year) when the crises of the 1990s hit. Such a rate would be a serious barrier to the developmental hopes of most countries, but especially one with the limited options of this country. Since then, however, population growth has fallen considerably, to an estimated 1.5â&#x20AC;&#x201C;1.9 percent by 2010.2 On the skills side there are a number of challenges, including that until the government fixes more clearly on a plan for the countryâ&#x20AC;&#x2122;s development over the coming decades, it is hard to guess what types of education and training will be the most in demand. Tajikistan faces at least two other major challenges to achieving a bright future. One is environmental. The cotton monoculture has been very destructive, and restructuring the large economically inefficient and environmentally damaging farms that date from the Soviet era is a major part of the answer. Other steps are also needed. Finally, the country has become the chief conduit for opium moving from East Asia to Russia, Western Europe and beyond and is, accordingly, subjected to the dangers associated with that industry, including corruption and addiction. Its recent history and current situation mean that Tajikistan presents many difficult challenges in the area of social protection policy. What is feasible and desirable is greatly affected by the limited institutional capacity to organize and deliver, and by the high level of emigration and the associated special problems and conditions. As a legacy of the Soviet system, Tajikistan has a non-contributory pension system directed mainly at old people, disabled people and survivors (Son, 2011: abstract). This system has greatly shrunk since 1992 and is not significantly progressive in its distributive effects, if progressive at all. The low incomes, the mainly rural and agricultural population and the weak institutional base from which to develop a reformed system of social protection all suggest that such protection may be most efficiently pursued on the production side or possibly by making the migration process more efficient. The seriousness of food insecurity highlights the importance of accelerating land distribution such that families can defend against it through their own production. Education presents a series of challenges and dilemmas. Health services have deteriorated seriously and need to be rebuilt; preventive interventions warrant attention in low-income countries such as Tajikistan given the evidence that they can generate very high returns on low costs (e.g. from provision of micronutrients, early detection of cancer etc.).

Tajikistan

161

The macroeconomic setting The drastic macroeconomic decline during 1988–1996 provides the background to the current challenges faced by the country. The precise nature of the current macroeconomic challenge, however, depends also on how one interprets the failure of the economy, judging by the official figures, to get back to its former peak level of output, much less approach the former peak level of GDP per capita even after 14 years of strong recovery. Various factors are or could be involved, including: i) Given the extremity of the economic crash caused by the dissolution of the Soviet Union and the ensuing civil war, it was possibly unavoidable that it would take decades to recover the lost ground. After a gradual acceleration between 1997 and 2000, growth (according to official statistics) then peaked over 2000–2005 at around 10 percent per year, before slowing to an average of 6.5 percent over 2006–2011. Two more years at that rate will bring official GDP back to the average for 1987–1989, but it will take 12 years or so to do the same for GDP per capita. ii) The huge loss of capital during the transition and the civil war lowered the potential output level of the economy such that only with considerable new investment can it much surpass the previous peak, especially in terms of per capita output. iii) The transition from a centrally planned towards a more market-oriented economy may have rendered some of the capital—physical, human and institutional—less productive under the new system, although it may also have raised the level of efficiency of resource use in other respects. iv) Serious mismeasurement of GDP and gross national income (GNI)3 may have exaggerated the economy’s initial decline and confused the picture with respect to its subsequent recovery. The failure to capture either informal-sector income or illegal income (drugs, other), whose shares of true GDP have been growing, has probably contributed to a considerable understatement both of the growth of that variable and of its current level relative to the previous peaks. In addition, another special consideration is that, with the last decade’s rapid growth of earnings abroad and resulting remittances, GNI has been increasing substantially faster than GDP, making the latter, even when correctly measured, a downward-biased indicator of income growth.4 With national accounts aggregates possibly very seriously understated and with inadequate data on how the components of the underground economy have been changing over time, it is clear that economic policy must be made using limited macroeconomic evidence. There is no reason to doubt that growth out of the extreme economic trough in which the country found itself by the late 1990s has been rapid, but now, with the underground economy possibly equal to or greater than the official one, it is possible that if the latter’s growth has slowed to zero, a reported GDP growth rate of 6 percent would in fact be only one of 3 percent, whereas if the underground economy continues to grow faster than the official one, that official 6 percent might really mean

162

Social protection, growth and employment

8–9 percent. Analysis of labour market and social protection issues cannot rely a great deal on macroeconomic data; thus conclusions on how well the growth process is contibuting to success or failure on the labour front must be based mainly on labour data alone. Fortunately, this is somewhat more reliable, despite some serious deficiencies.

International trade Tajikistan is fairly typical of very small countries in having only a couple of products that dominate total export revenue, and a low share of world markets in each—less than 4 percent for cotton and less than 5 percent for aluminium. Promising new exports include hydroelectric power, where Tajikistan ranks eighth in the world in terms of exploitable resources, with 4 percent of the world’s cost-effective potential. Like other small open economies, Tajikistan is strongly influenced by cycles of the world economy. The fall in world market prices for aluminium and cotton during the world crisis led to a trade deficit that in recent years reached 35 percent of GDP, essentially the other side of the coin from the large net inflow of remittances. The problem of price shocks in the world markets for exported raw materials and in the flow of remittances is likely to remain for many years.

Savings and investment According to official figures, gross fixed investment accounted for 14 percent of 2005 GDP; about two thirds of it was public. Subsequently this ratio rose to a peak of 22 percent in 2007 before slipping back to 19 percent in 2010. Private investment is of particular concern; although its share of GDP has risen since a trough around 2000, the average over 2007–2010 has been only 6 percent of GDP, far too low to be the basis for a sustained level of growth. Part of the understatement of GDP (e.g. failure to capture most informal activity) referred to above is presumably matched by a perhaps comparable understatement of private investment.5 Further, some expenditures which in reality constitute a form of investment are not included, most prominently expenditures undertaken in connection with emigration (see Section 3).

The financial system and financial intermediation Formal credit markets remain very underdeveloped in Tajikistan, not surprising given the recent and still partial transition towards a market economy and the low levels of savings in the formal economy, leading to low deposits in the financial intermediaries. Allocation of credit to private businesses is correspondingly low, at 5 percent of GDP, and very few funds find their way to start-ups and small and medium-sized enterprises (SMEs). Among factors cited by businesses in describing the difficulty of obtaining loans, the most prevalent are collateral requirements and unofficial costs (bribes). In 2005, the collateral requirements for a loan were almost twice the principle, and unofficial payments were estimated at 10–15 percent of the loan value.

Tajikistan

163

The role of remittances With earnings in Tajikistan being low, the external labour market (primarily jobs in Russia) offers a way around the shortage of decently renumerated jobs. The average wage earned by migrant workers in Russia is far above domestic levels; as of the early 2000s the net private return to an investment in migration to work temporarily in Russia was estimated at about US$150 a month (World Bank, 2005), making it highly attractive and giving Tajikistan one of the highest ratios of emigrant remittances to GDP. To evolve into something more than a supplier of mainly labour to Russia and perhaps elsewhere, the country clearly needs to create a much stronger domestic demand for labour. In the meantime, emigration and the accompanying remittances serve the very useful safety valve function of providing remunerative jobs and supporting family incomes. At present this can be seen as by far the most important form of social protection in Tajikistan, albeit unplanned and not without its own problems. The presence of remittances on this scale raises two broad questions: i) Can they help to leverage a growth take-off, while still performing their function as stopgap to prevent incomes from falling too low?; and ii) What types of social protection and other policies can help to make the emigration/ remittance process more efficient and successful, while assuring that its social costs are minimized? To deal with these questions, one must look in some detail at the migration phenomenon in Tajikistan, its character, scope, trends and effects. In the last decade or so a huge wave of mainly temporary emigrants from Tajikistan have made their way to Russia to seek work, mainly in response to the economic hardships discussed above. A command of the Russian language, visa-free entry to Russia and well-established migrant networks lower the barriers to and the costs of migration. There are still problems; most migrants are considered illegal, since they lack work and residence papers, and they are usually employed in â&#x20AC;&#x2DC;3Dâ&#x20AC;&#x2122; work (Dirty, Dangerous and Disliked).6 There have been reports of migrants being deprived of basic civil and worker rights in Russia, and some have been subjected to exploitative parctices by employers and officials (Olimova, 2003). The lack of legal status makes them vulnerable to rent-seeking by corrupt officials; one survey (Jomnes et al., 2007: 20) found that they spend an average of US$165 on bribes and various unspecified fees. All such barriers notwithstanding, there was a huge increase in emigration between 2004 and 2008, and remittances jumped from a few hundred million US dollars to over US$2.5 billion (ILO, 2010: 55). Recent estimates have put the number of Tajiks working abroad at a given time at between 600,000 and 1 million,7 with approximately 95 percent of them in Russia. Even the lower figure represents almost 15 percent of Tajikistanâ&#x20AC;&#x2122;s total working-age population, 30 percent of the males in that category, and an even higher share of young males. The corresponding remittance flows have been estimated at anywhere between 30 and 50 percent of official GDP;8 a

164

Social protection, growth and employment

recent figure from the National Bank of Tajikistan for 2011 is US$2.96 billion or 45.4 percent of Tajikistanâ&#x20AC;&#x2122;s official GDP.9 Dependence on remittances is essential to the livelihood strategies of many Tajik households. As they rose rapidly during most of the decade begining in 2000, their contributions and effects no doubt also changed. By 2008 an International Labour Organization (ILO) household survey found that over 70 percent of families had at least one member abroad (29 percent had two or more) and that for 62 percent of households remittances accounted for 60 percent or more of income, with only 11 percent having no income of this sort (ILO, 2010: 12, 24). According to an earlier study (Brown et al., 2008: 11), which reported that up to 1 million people or 15 percent of the population lived in households whose main source of income was remittances, 37.3 percent of all households had at least one migrant member, with this figure rising to 42.3 percent for rural areas. In households that received them, remittances accounted on average for 59.8 percent of total income. Extreme poverty was found to be significantly less (18.8 percent) among households that received remittances than on average (32.2 percent); in the absence of remittances the incidence of poverty would no doubt have increased a great deal.10 Remittances probably also help to reduce inequality,11 which appears to be relatively high in Tajikistan, especially in the capital.12 Data from some years earlier (1999, it appears) indicated that remittances were important for all quintiles of families ranked by total household income but accounted for only 10 percent of that total income (although about 17 percent ot earned income excluding value of food produced, gifts and social assistance) (Falkingham and Klytchnikova, 2006: 18). Presumably these shares rose significantly as the decade proceeded. Under some conditions emigration/remittances also lower income insecurity, especially when jobs elsewhere constitute a sort of large and stable shelf of possibilities. Jones et al. (2007) note, though, that even for reducing poverty, migration can be an unreliable mechanism. For Tajikistan, where both income and food insecurity are major problems and challenges, that vulnerability is a serious cause for concern.13 As a medium-term economic strategy, migration has drawbacks at both the family and the national levels. A householdâ&#x20AC;&#x2122;s decision to send a migrant abroad implies costs and risks, which are weighed against the prospects of generating income locally. The cost of migration to Russia (estimated in 2005 at US$410 on average) (Mughal, 2007: 101) probably renders it difficult or impossible for some poor families; given the infrequency of female migration, it is also beyond the capability of families without an eligible male who can emigrate. At the national level the great dependence on remittance income has become a significant political concern, since its sustainability depends on factors outside the countryâ&#x20AC;&#x2122;s control, and the macroeconomic effects of a serious barrier to Tajik migration to Russia would be dramatic. In November 2010, amid tensions surrounding the arrest of two Russian pilots in Tajikistan for alleged smuggling, the Russian government withdrew its ambassador and began to round up Tajik migrants for deportation under the pretext that they posed a public health concern to the local population.14 Apart from vulnerability to Russian political influence, dependence on this income source exposes

Tajikistan

165

Tajikistan to the the hydrocarbon-dependent business cycles in Russia. The effects of a downturn in Russia are a serious cause for concern in Tajikistan on this count as well. Migration of the current magnitude naturally has major effects on the domestic labour market; the phenomenon reduces domestic labour supply, and the potentail for it raises the wage level at which Tajiks are willing to take a local job (the reservation wage), induces some to remain unemployed instead of accepting a low-wage job and lowers the level of informality, as argued by Abdulloev et al. (2011).15 A tighter labour market increases the wage bill faced by employers in Tajikistan and, together with a lack of people in some skill areas, can pose an obstacle to business success in some areas. Rising wages in response to labour market tightness are in general desirable, but when an economy is striving to build a base of private enterprise that will eventually provide better employment options, such increases can be premature if they slow the process of enterprise development that contributes to long-term growth. Other specific concerns about emigration/remittances relate to their joint impacts on education and health (see below) and on child welfare in a more general sense.16 By one estimate, more than 9000 children whose parents have migrated are left without supervision, exposed to child labour and not in school. Although non-governmental and international organizations have made efforts to address this problem, without organized government support it is difficult to achieve much progress.17 Emigration is sometimes also blamed by some for further fracturing an already disjointed society. A longer-term concern involves the risk that, if the country’s economic prospects do not improve, especially in the agricultural sector, the export of labour will evolve into a permanent loss of human resources. The ILO reports that, despite its generally temporary nature, half of all emigrants are ‘long-term’, as defined by being away for over two years with either no or only brief trips back (ILO, 2010: 16).18 If more and more migrants become less temporary, remittances will eventually fall, and their contribution to Tajikistan’s current welfare and future economic development will be reduced. One risk is that temporary emigration may, by taking pressure off the system, pre-empt a strong effort by the Government of Tajikistan to create new jobs, mainly by supporting private enterprise, especially family agriculture and SMEs elsewhere in the economy. That emigration and remittances have been desirable and indeed essential mechanisms of survival does not render this longer-term policy question irrelevant. The international literature provides some limited guidance on it, if not with respect to clear conclusions, certainly with regard to questions to be asked and considered.

Emigration/remittances as a motor of economic growth? When can migration revenues become a motor for growth? Studies of other countries that receive a high level of remittances have clarified some of the economic impacts and the potential to contribute to growth.19 Beyond spending for family sustenance, considerable flows are in some cases directed at (i) improving housing and other longer-term assets; (ii) establishing businesses; and even (iii) contributing to the local community from which the emigrant has come (notable

166

Social protection, growth and employment

in Mexico). All of these uses would be welcome in Tajikistan, since all could contribute to the future welfare of the population.20 The evidence available thus far, though not very encouraging with respect to its likely direct contribution to private-sector and overall economic growth, does have some promising elements. At one extreme, Brown et al. (2008: iii) report that 98 percent of remittance income is used to cover expenditures on food and other non-durable goods and services (including rituals such as marriages); their use is almost entirely non-business in character, with only 1.5 percent directed to private business, luxury goods or durables. Even for that small amount of remittance revenue going into business investment, there is no clear evidence that it has a positive effect. In fact, the study reports that households receiving remittances invest only half as much on business expansion as non-receiving households do. This is likely due in part to the gender structure of migration, since remittance-receiving households tend to lack males with an entreprenurial bent, together with the probability that a lack of investment opportunities is often a factor contributing to the migration in the first place. The ILO (2010: 26, 36), drawing on a different survey, reports that 57 percent of remittances are used for immediate consumption needs, 12 percent saved for six months or less, and 11 percent saved for longer periods. The reasons given by the emigrants for seeking the income or for saving do not suggest much eventual business investment. Only 2 percent of respondents receiving remittances reported having an active bank account, and this seemed unrelated to whether they saved or not. The principal reasons for saving are ‘rainy day’, house purchase/maintenance, events, health care and tuition, with the category ‘other’ (which presumably includes investment) coming in at 8 percent. Only 5 percent of respondents were aware of saving schemes offered by the financial sector. More promising is the fact that 12 percent of households with family members working abroad are engaged in a business activity (ILO, 2010: 41). The main types are retail commerce, at 68 percent of the total, transport (10 percent), agriculture (9 percent) and services (6 percent); 52 percent of the businesses involve just one person, 47 percent involve other family members as well, and just 1 percent have other partners. Of these households with emigrants working abroad, 30 percent were interested in continuing or starting a business; given that 12 percent were engaged in a business, this implies that 18 percent were interested in starting one. Several cross-country studies have pointed to promising benefits from remittances. Vaaler (2011: 1121), drawing on data from 61 developing countries over 2002–2007, concludes that remittances increase the availability of venture capital; they also increase business start-up rates in countries where the public sector is sufficiently small (and presumably less of an obstacle to business). The apparently very limited channelling of Tajikistani remittances into future productive assets (except for the part going to education) is unsurprising in light of the very low incomes of most receiving famlies and the still diffucult context for private business. But it is important to know whether it is also in part due to a lack of entrepreneurial spirit, experience or talent (all understandable in an economy in transition) and/or to the barrier to new opportunities posed by the market power of the inner circle that controls much of the economy. It might also be that the presence of drug trafficking discourages other activities.21

Tajikistan

167

The domestic labour market, incomes and poverty For many years Tajikistan’s labour market has been in a state of excess supply, leading to the massive exodus to Russia. Population growth reached a dramatically high rate of over 3 percent during the late 1980s, after which it fell markedly by the early 2000s, due to the combination of a high death rate, emigration and falling family size, and then climbed back by 2010 to somewhere between 1.4 and 1.9 percent. The average annual net growth of the working-age population was around 4 percent over 1996–2005, due to a rising ratio of working-age people to total population — a lagged effect of the earlier very fast population growth and a pattern that is likely to continue for some time, thereby keeping supply pressure on the labour market. Over 1991–1997 average labour productivity declined precipitously as the economy plunged; recovery since then has been marked but still far from sufficient. Productively absorbing the new entrants into the labour market will require continued fast and reasonably labour-absorbing economic growth for a sustained period. While a member of the Soviet Union, Tajikistan had an essentially formal labour market, where the state was responsible for ensuring that everyone had a job.22 Since then the public/ state sector has shrunk, due in part to the transition to a more market-oriented system and also to the overall economic collapse brought about by the decline of the Soviet economy (to which Tajikistan remained substantially tied) and the ensuing civil war. Real wages fell drastically, as inflation soared. According to official figures23 for 2000, the public and collective sectors (the latter in agriculture) provided about 56 percent of jobs; by 2007 this share had fallen to 47 percent and by 2011 probably to about 44 percent. Employment in the private sector has grown from 19 percent of the total around independence to an estimated 43 percent in 2000, 52 percent in 2007 and probably about 55 percent by 2011. Since very little private-sector activity is ‘formalized’ yet, in the sense of abiding fully by the sort of tax, labour and other legislation that prevails in the larger-scale private firms in market-oriented economies, the great bulk of this employment is appropriately defined as informal.24 By the first decade of the present millennium, with domestic formal employment having fallen in both absolute and relative terms, nearly half of jobs were informal, and the number of workers emigrating to Russia was on the increase. Most estimates for recent years tend to put informal jobs at close to 50 percent of the total,25 just a few percentage points below private-sector employment as a whole. The increase in informal activities over the post-Soviet period has been related to the outflow of workers from the formal sector, the slow growth of private, formal employment and the growth of open unemployment. The informal sector acts as a large labour reserve during long-term transition processes like that through which Tajikistan has been proceeding, and as a shorter-term shock-absorber that expands in times of economic recession and shrinks during periods of economic recovery. Open unemployment has been a significant though not dominant phenomenon,26 at 7.4 percent of the labour force according to the Living Standards Survey of 2004, and tends to be quite long-term, with nearly half searching for over a year and less than a quarter for under three months (Babajanov and Khakimov, 2010: 36, Table 17). According to

168

Social protection, growth and employment

that survey, 70 percent of the total number of unemployed young people were actively looking for work, while 30 percent were trying to start their own business; only 5.5 percent had recourse to the public employment service, and 7.2 percent to the labour exchange. The transfer of workers from formal to informal activities has been partly voluntary and partly not. One factor has been forced leave (without pay or with partial pay) at the initiative of enterprise management. In 2008, such individuals constituted around 3 percent of the full staff of those enterprises, but presumably the possibility of this outcome encouraged others to leave as well. The transition path in Tajikistan’s labour force appears, at least in qualitative terms, to have followed the typical pattern in formerly centrally planned economies, as described by Dimova et al. (2006).27 As attested by its size, the informal sector in Tajikistan often provides a family’s only or best alternative to unemployment; for many others it is the main alternative to emigration. Young people of working age have increasingly resorted to such employment, thereby sacrificing the unclear prospects of acquiring the pension rights traditionally associated with formal employment28 and the currently even less clear rights to other government social services for the more regular and in many cases higher earnings that informal employment provides, despite its own limitations, including income instability due to the severe demand fluctuations faced by many small enterprises. Informal employment consists primarily of people engaged in their own agricultural family enterprises or outside agriculture in private microenterprises involved mainly in trade, services and construction, together with paid workers and unpaid family workers in smallish farms or commercial enterprises. A substantial number of formal-sector workers are also forced to work part-time in the informal sector to generate additional income. By one estimate (cited by Babajanov and Kakamov, 2010: 65), wages from formal employment constitute only one third of total household income, the remainder coming from the informal agricultural activities, remittances from migrant workers and some government subsidies.29 As in other countries, informal employment often provides a part-time option for students and pensioners, who now make up about one sixth of all informal workers. Finally, as noted in Section 6.2 below, informal employment sometimes provides entrepreneurial experience that pays off increasingly over time.30 The traditional downsides of informal employment are that it does not meet the labour and social guarantees of most formal employment (social security, pensions etc.), usually provides only modest incomes, and in many cases gives little chance of professional development. The very limited tax revenues generated from these activities constrain public budgets at all levels, contributing to a shortage of domestic investment and underfunded social services. Intensive use of child labour in seasonal agriculture adversely affects the development of human capital in the country. Given the above complexities in how and where labour is applied, wage levels and trends are hard to estimate. Babajanov and Kakamov (2010: 55) note that official salaries in the private sector are currently higher than in the state sector, but that only 25–30 percent of the wage is formalized, with the remainder paid as part of an off-the-books surcharge; this component of the wage does not provide social security benefits. In many companies a maximum of 25 percent of employees are ‘officially’ registered as being employed.

Tajikistan

169

Table 4.1: Sources of Cash Income, Tajikistani Households, 2006 Source of income

Percentage of total cash income

Remuneration from a state-owned institution

15.2

Remuneration from a kolkhoz (collective farm) or sovkhov (state farm)

11.6

Remuneration, private farm, not belonging to the family

1.3

Remuneration, other private business

8.1

Cash income from farming activities

24.4

Cash income from other activities

17.4

Cash transfers from the state Cash income from household members abroad

9.7 10.6

Other

1.6

Total

100.0

Source: Olimov (2007: 55). The period to which the data refer is not made clear, but the survey was undertaken in 2006 (ibid.: 8).

Table 4.1 provides one estimate of the sources of work-based cash income of Tajikistani families in 2006. Meanwhile, Olimov’s (2007: 8) data on total income indicate that food for home consumption amounts to 14.7 percent of official GDP, and wages paid in kind and barter exchanges to 13.2 percent.31 Olimov (ibid.: 9) estimates home production of goods and services for consumption and sale at 23.3 percent of total income, with food products at 14.8 percent of GDP.32 As would be expected in light of the fast growth and the surge of emigration and remittances since the late 1990s, there appears to have been a large decline in poverty over that period. According to the Living Standards Surveys of 2003 and 2007, the percentage of families without income adequate to satisfy basic needs fell from 72.4 to 53.5 percent (Babajanov and Khakimov, 2010: 12). Poverty incidence does not differ much between urban and rural areas, being only slightly higher in the latter; as of 2007 the level of absolute poverty in urban areas (18.9 percent) was actually higher than in rural areas (16.4 percent). The fact that poverty is still widespread is reflected in the high (73.7 percent) share of total household expenditures going to food.33 Presumably poverty remains well above the levels of the later Soviet years. The level of poverty reflects low wages in general. Those of women are especially low, with average earnings far below those of men—as of 2000 the ratio was 43.2 percent according to official statistics; by 2006 it had reached 55.3 percent. One major source of the gap is gender segregation of employment. The highest proportion of women are found in agriculture, health care, education and culture, arts and sciences—the four sectors where wages and salaries are below the economy-wide average. Even in agriculture, with the lowest average earnings, women earn almost one third less than men. In low-income sectors with highly qualified personnel (education, health and culture) women’s earnings are again well below men’s34 The two sectors where women’s wages do not differ significantly from those of men—public administration and financing and sectors of material production—employ only 4.2 percent and 0.8 percent of the

170

Social protection, growth and employment

total, respectively, and have low female shares of employment (28.5 percent and 17.9 percent, respectively). In light of these gaps one may guess that an important factor in encouraging women to enter informal work is the income earned. In 2008 the average nominal monthly wages and salaries for women in the informal sector was TJS235.4—61 percent of the figure for men (TJS383.8).35 Reflecting the urgent need of some low-income women for employment is the organized informal labour market for women called mardikorov. This market connects employers needing workers to do traditionally female types of jobs (such as weeding, picking fruits and vegetables, canning or pickling, washing clothes or dishes after any major events, house cleaning or guest services) for very short periods of time, with terms of work established case by case. Some jobs are carried out by individuals, others by groups that may be pre-organized or may form at the market based on the needs of employers. A teamwork form of labour organization brings together women living in a village and headed formally by a female ‘Brigadier’. Part of this spontaneous female labour market is seasonal. The share of all able-bodied women involved in this market is low, perhaps 0.5–1.0 per cent. Most are from very poor families—people who, faced with more attractive options, would not have entered this market. Some are widows, abandoned by their husbands or having three or four dependents and unmarried. Some with complete secondary education or even higher education, after failing to find jobs corresponding to their skills, have recourse to the mardikorstvo. Other poor women with an equally urgent need for jobs engage in prostitution, fraud, drug trafficking etc. Although informally employed mardikorstva workers can, depending on the complexity and harshness of work, earn around TJS25–60 per day, considerably higher than the wages of men and women employed in the formal sector, due to the short-term and unstable character of the work, the earnings are usually only enough to feed the family. The situation of child labour in Tajikistan is disturbing, consistent with the high incidence of poverty and the absence of the many young men working abroad. According to the UNICEF Multiple Indicator Cluster Survey (MICS) of 2005, 200,000 children aged five to 14 were engaged in child labour (excluding non-intensive household chores), with 65,000 of these in paid work. Most were attending school, but 20,000 were not, because of their work (Baschieri and Falkingham, 2007: 7–8). Child labour was more frequent among those in the poorest households (15.6 percent) than among those in the richest ones (9.5 percent). The international fund Save the Children (2008) reveals that Tajikistan has the worst forms of child labour, especially in the agricultural sector.

Structural problems of the Tajikistani economy: the food and energy crises Despite the many areas of concern, the period between 2000 and 2007 was one of considerable advance and recovery. The economy managed to get closer to pre-independence levels of output in important areas such as agriculture. But the multiple crises that have shocked the world since 2007/08 have once again highlighted economic vulnerability, even and in some cases especially in remote and isolated nations such as Tajikistan.

Tajikistan

171

As the global financial crisis hit the Tajik economy through its effects on remittance flows from migrant workers in Russia, one of its main consequences was to exacerbate an already looming food security crisis. The official growth rate dropped from 7 percent in 2008 to 3 percent in 2009, external remittances dropped from an estimated US$2.6 billion to US$1.6 billion, and there was a substantial fall in foreign direct investment and humanitarian assistance. Compounding these woes, Tajikistan faced—and still faces—a crippling energy and fuel crisis. Oil exports from Russia have recently been subjected to export taxes, and completely banned for some periods. Climate change is lowering water levels in the river systems, a potentially catastrophic scenario for a country that relies entirely on hydropower for its energy supply. An extraordinarily severe winter in 2007/08 sparked a calamitous energy crisis during 2008 and brought into focus the fragility of Tajikistan’s water–energy nexus.

Food insecurity While international food price spikes such as that of 2008 are not unprecedented, several fundamental differences between that crisis and previous ones have implications for nations such as Tajikistan. One involves the likelihood that real agricultural prices will not soon resume their decades-long downward trend but will more likely stabilize or rise due to changes on both the demand and supply sides of the market. Demand pressure from a growing global population (surpassing 7 billion) is now coupled with increasing incomes in many developing countries, creating a burgeoning demand for meat, vegetable oils and other high-end food products and putting pressure on the agricultural land base. The introduction of biofuels as an alternative energy source has created a new and competing use for food crops.36 On the supply side, the consistent decline in investment in agriculture in developing countries and, most critically, in research and development, and the associated deceleration of productivity growth, is a defining difference from previous decades (Timmer, 2009). In countries such as Tajikistan the increase in food prices is not at present a sufficient impetus to such investment; lack of rural infrastructure, limited access to modern inputs and irrigation, poor roads and storage facilities, rudimentary technology, limited knowledge of modern farming techniques and limited access to credit all contribute to low productivity, limited participation in markets and lack of investment. In Tajikistan, as in most low-income food-deficit countries, the structural solution to the problem of food insecurity lies mainly in increasing the production and productivity of small family farmers. Regardless of the longer-term agricultural output and price trajectories, it is generally agreed that the world has entered an era of food price volatility due to a combination of the increasing role of financial markets in price determination, new links between agriculture and energy markets, and changes in agricultural trade policies that have led some developing countries to rely more heavily on trade in staples and others to move towards self-sufficiency (Naylor and Falcon, 2010). Energy-related inputs (fertilizers, pesticides, and fuel for farm machinery and transportation) represent a larger component of agricultural costs than before, and petroleum prices serve as a reference point for the profitability of maize-based ethanol and hence affect the demand and price of maize. With the unpredictability of today’s energy markets, the increasing links to and dependence on them will be a source of turbulence in international agriculture

172

Social protection, growth and employment

Figure 4.1: FAO Food Price Index in Real Terms, 1961–2010

Notes: Calculated using International prices for cereal, oilseeds, meats, dairy products and sugar. The FAO Food Price Index is calculated from 1990 to the present on a regular basis; in this figure it has been extended back to 1961 using proxy price information. The index measures movements in international prices and not necessarily domestic prices. The United States GDP deflator is used to express the Food Price Index in real rather than nominal terms. Source: Calculations by FAO.

markets. Furthermore, climate change may lead to more frequent and extreme weather events and the consequent shocks on agricultural markets from the supply side. Real international food prices have been increasing since the early 2000s (Figure 4.1), culminating in the price spike of 2007–08. Although they dipped in 2009, they have remained well above their trough of around 2000 and reached an even higher peak in early 2011 (Figure 4.2). In the lead-in to the recent spikes, the trend in Tajikistan follows the global pattern, albeit with less volatility. Unfortunately, just as Tajikistani food prices began to increase during 2008, the flow of remittances began to decline (Figure 4.3), putting a double strain on household budgets. When international food prices dropped sharply after July 2008, the local prices in Tajikistan remained near their peak levels and have increased further as the second upswing in world food prices took hold in 2010. Over 2007–2011 the net increase in Tajikistan has been very close to the global increase. Food insecurity has been a feature of the Tajik economy from well before the recent grain price spikes, so the current global context exacerbates a chronically precarious condition. The

Tajikistan

173

Figure 4.2: Food Price Trends in Tajikistan and Globally, 2007–2010 180

Global

170

Tajikistan

160 150 140 130 120 110

May-11

Jan-11

Mar-11

Nov-10

Jul-10

Sep-10

May-10

Jan-10

Mar-10

Nov-09

Jul-09

Sep-09

May-09

Jan-09

Mar-09

Nov-08

Jul-08

Sep-08

May-08

Jan-08

Mar-08

Nov-07

Jul-07

Sep-07

May-07

Jan-07

Mar-07

100

Source: State Statistical Office of Tajikistan, FAO; UNDP calculations.

Figure 4.3: Remittances for Tajikistan 2003-20011 (US$ Billions)

Source: World Development Indicators Database, World Bank

UN Food and Agriculture Organization (FAO) (2010) categorizes Tajikistan as one of 22 countries in a protracted food crisis, based on three criteria: i) having been faced by crises (whether human-induced or natural) for an extended period of time –Tajikistan has experienced 11 such crises between 1996 and 2010; ii) more than 10 percent of the country’s overall official development assistance (ODA) has taken the form of humanitarian aid since 2000 –13 percent for Tajikistan between 2000 and 2008; and

174

Social protection, growth and employment iii) being classified as a Low Income Food Deficit Country—between 2005 and 2007, 30 percent of Tajikistan’s population was deemed to be malnourished; for the 22 countries in question the average rate of malnourishment was 37 percent; ranging from 14 percent (Ivory Coast) to 69 percent (Democratic Republic of Congo). By contrast, China and India had rates of 15 percent, and the figure for all other less developed countries taken together was 14 percent.

Tajikistan faces serious natural constraints to food production. Its high, mountainous terrain leaves limited space for cropping—just 0.14 ha/person compared to the global average of 0.26 ha/person. The lack of good land has led to expansion into more marginal areas, heavy use of irrigation, and serious environmental damage in the form of land degradation, soil erosion, flooding, salinization and desertification.37 An additional threat to food security is climate change. Temperature increases in mountainous Central Asia are predicted to reduce crop yields by 2.5–10 percent by 2020 and by 5–30 percent by 2050 (IPCC, 2007). The increased heat stress on crops and grasslands during summer months will increase the demand for irrigation. Higher frequencies of extreme weather conditions will also have an adverse impact. The effects of climate change on river flows are perhaps the most worrying concern. In the long term, as many glaciers in the mountains of Tajikistan shrink in size and disappear, the water flow originating in glacial melting will be reduced, leading to a significant long-term decrease in major river basins, with very serious potential implications for both agriculture and energy production.

The energy challenge Electricity generation in Tajikistan is based almost entirely on hydroelectric stations, with nearly 70 percent of total capacity located at the Nurek power plant on the Amu Darya, east of the capital, Dushanbe. Tajikistan has a very large hydropower potential and currently uses no more than 5–10 percent of it. It thus remains a key sector for economic expansion, with several internationally financed dam projects, including Sangtuda and Zeravshan, currently underway. Technical challenges include an erratic pattern of runoff in some river basins that can have significant effects on power output. The importance of energy to Tajikistan is hard to overstate. It provides a key input to the important aluminium sector and to agriculture, as well as being central to the general health and well-being of Tajik citizens. Many of the towns and villages in the mountains and foothills are located above the watercourses. Prior to the transition, electrically powered water pumps raised water from the rivers to these settlements. Since then the periodic energy deficiencies have created shortages of drinking water, affecting even Dushanbe, where about 25,000 people in the eastern foothills of the city remain without a regular supply. Irregular pumping leads to stagnant water in reservoirs, forcing many rural residents to resort to irrigation ditches and as a result to become victims of such waterborne diseases as Hepatitis A and E and Typhoid Fever, leading causes of mortality in both adults and children. In the absence of electricity, residents have to build canals or irrigation ditches from lateral gorges located uphill to draw water using gravity; this requires considerable investment of material resources, labour and time. The irregular

Tajikistan

175

supply of electricity, heating and water also disrupts the services of health centres and educational institutions. Faced with power shortages, households are also pushed towards alternative sources of energy for heating and cooking. The felling of trees and shrubs has led to worrisome declines in the coverage and volume of forests and at the same time has increased the risks of soil erosion and landslides.38 Apart from firewood, the most readily available source of energy is biomass and animal waste; however, burning these products for heat and energy raises serious health concerns and diminishes the supply of organic fertilizer, which, given the difficulties in obtaining artificial fertilizers, is a critical ingredient to maintaining the natural fertility of the soil. Success in the energy policy area is complicated in part by technical aspects. Water management is now a hugely politicized issue in the region, as upstream water withdrawals for irrigation and the creation of reservoirs along the Amu Darya have significantly reduced the flow reaching Turkmenistan and Uzbekistan. Tajikistan’s future water management decisions, including the construction of dams for hydroelectric generation and further diversions for irrigation, are thus not only a major issue locally but also internationally. Central Asia’s energy currently comes primarily from large hydropower dams in Tajikistan and the Kyrgyz Republic. When the region was under centralized Soviet management, the Torktogul reservoir in the Krygyz SSR and the Nurek reservoir in the Tajik SSR were primarily used as reservoirs for summer irrigation for the water-intensive agricultural sectors in the downstream Uzbek and Turkmen SSRs; in return these SSRs provided subsidized fossil fuels to their upstream neighbours to generate heat and electricity in winter. When the post-Soviet oil and gas systems were marketized after the independence of these CIS countries, the exponentially higher prices for gas and fossil fuels forced the Kyrgyz Republic and Tajikistan to switch to using their reservoirs as hydropower generators. Much of the conflict between these countries since then has focused around this water–energy nexus. The importance of energy development to Tajikistan and the regional complications were both underlined by the calamitous energy crisis during the year after the exceptionally cold winter of 2007/08. Together with a drop in water inflow to the reservoirs came a drastic increase in its release to meet the huge surge in energy demand; this combination pushed the water levels in the dams dangerously close to the ‘dead level’ at which energy generation is no longer possible. To make things worse, the unprecedented winter water release caused flooding in the Syr-Darya basin and damaged the downstream countries’ water infrastructure. The crisis was further exacerbated by drought conditions in the Aral Sea Delta during the following spring and summer when Uzbekistan and Turkmenistan needed water for their vast cotton fields. This sequence highlighted the vulnerability of the region’s energy sector and its ancillary effects on food production, and underlined the sort of effects that climate change can have on such a fragile ecosystem. A reliable energy supply is pivotal to the industrial and agricultural sectors of Tajikistan, which together account for 80 percent of domestic electricity consumption. Aluminium, the country’s single most important export, relies on Tajikistan’s ability to generate reasonably cheap hydroelectricity, and disruptions in power supply to it have a direct impact on Tajikistan’s GDP. The agricultural irrigation system relies on electric pumps and electricity, and the water that it

176

Social protection, growth and employment

provides to villages is, as noted above, directly important to the quality of life. It is, therefore, urgent not only to reverse the decline in energy supply but also to increase it above previous levels. Energy can provide a catalytic impulse to spur economic activity in Tajikistan and, managed properly, could help to achieve a pro-poor pattern of growth. Food security, domestic investment that relies on a good business environment, and agricultural productivity all stand to benefit from a concerted effort in the energy sector.

Family agriculture and small-scale non-agricultural enterprise Agriculture and small non-agricultural private enterprise are the two sectors that must generate the bulk of employment in Tajikistan for at least the next couple of decades. Successful development of small-scale family agriculture is especially critical to achieving improved food security and reducing poverty. This section reviews progress to date and the potential and challenges facing these two sectors.

Family agriculture and the cotton sector Agriculture remains the main activity in the domestic Tajik economy, with about half of all employment and a fifth of output. Unhappily, the sector also houses the bulk of the countryâ&#x20AC;&#x2122;s poor and extremely poor people. The Soviet era saw the Central Asian region become a leading exporter of cotton on the basis of extensive irrigation networks, a legacy that the economies still try to emulate.39 Farmland was collectivized and organized into state farms (sovkhoves) and collective farms (kolkhozes), the former managed directly by the government, and the latter by an elected administration from the farm and with approval by the regional Party committee. Targets were set by Moscow. Both types of farm were very large â&#x20AC;&#x201D; typically 1000 ha or more â&#x20AC;&#x201D; and were a principal unit of social organization in rural Tajikistan. Each family was provided with a house and adjacent plot (kitchen garden) to grow food for household consumption; in return the family worked on the large farm, whose management paid their salaries and was responsible for providing health care, education and social welfare. While production flourished, the monoculture had drastic ecological effects, including widespread desertification, salination and, most notably, the drying up of the Aral Sea. Inputs, including fertilizer and seed, were obtained through the command system. At independence the new nation states no longer had automatic access to these inputs; by virtue of its isolation and the lack of capacity to supply them, Tajikistan fared the worst; Uzbekistan, in contrast, was a major fertilizer-producing republic and thus better equipped to make the transition. Output from the kitchen gardens led to the first private markets under the Soviet system, as households began producing more than their own needs and selling the surplus in local markets

Tajikistan

177

and along the roads. By the end of the Soviet era, up to a third of the food sold in those markets came from kitchen gardens. When the country declared independence from the USSR in 1991and soon thereafter descended into civil war, they became the primary source of survival for the Tajik population. They continue to be the most productively utilized land in the country. With Tajikistan’s agricultural sector under the Soviet system not geared towards food, independence led to a severe supply crisis. In 1992 the government passed its first land reform package (the law ‘On Land Reform’), aimed at alleviating that crisis. It designated a special category of ‘presidential land’ from the excess not already under sovkhoz/kolkhoz management; households could draw on it to expand their kitchen gardens. In 1996, as the country began to emerge from the civil war and with the farming sector in deep financial crisis, it decreed a more comprehensive reform that aimed to reorganize the kolkhozes/sovkhozes into dehkan farms (dehkan is the Tajik word for ‘farmer’). Although the land in dehkan farms remains the property of the state and cannot be bought or sold, the farmer is granted inheritable rights to its use and the freedom to manage it as he sees fit; the state reserves the right to take back this land at any time if it is not being used ‘effectively’. An April 2001 law ‘On Land Use Planning’ laid the foundations for registration and titling of land but again retained strong state powers to intervene, under the guise of land use planning. Land was scheduled to be completely transferred into this dehkan system by 2005, with the exception of farms set aside for seed production, livestock breeding and research. In fact the process was slower, but by 2006 the agricultural land in ‘corporate farms’ (the successors of former collective and state farms) had shrunk from the Soviet level of 99 percent of the total to 35 percent. Most of this land had shifted to the dehkan category, which now accounted for close to 60 percent, with the remaining 5–6 percent in household plots, up from their traditional 1 percent in the Soviet era. In fact, however, the reforms have not produced anything approaching a normal family farm system. On the one hand, the government simply converted kolkhozes/sovkhozes into corporate farms (now called ‘collective dehkan’), the apparent changes thus being essentially cosmetic in nature. The transfer of land to the tenants, the critical aspect of the reform, was to be achieved through a process of share allocation formalized through membership certificates, but this process has been very slow and partial. Only 5.6 percent of farm workers interviewed by Action Against Hunger (AAH) in 2003 reported having received these certificates (Porteous, 2003: 15–19); most were completely unaware of the changes, with 64.3 percent believing they were still working for a kolkhoz/sovkhoz. The other category of farm emerging from the reforms is the ‘independent dehkan farm’, typically less than 50 ha, run by an individual, a family or a group of families and formed through the initiative of individual farmers in one of two ways: a shareholder or a group of shareholders of a large farm can apply to the farm management and district hukumat (governing body) to withdraw their shares and use them to start their own farm; alternatively, anyone can apply directly for land from the special fund of unused land of the district. The value of the latter option is greatly reduced because the land is typically of low quality and rain-fed rather than irrigated. While this independent dehkan farm is the stated aim of the land reform, the procedure of ‘privatization by application’ has greatly impeded the process. Those who are familiar with the laws, have personal connections with local authorities and can

178

Social protection, growth and employment

Table 4.2: Estimates of the Share of Individual and Corporate Sector in Land, Based on Data from Goskomstat and the State Agency for Surveying, Cartography and Land Use, 2006 Agricultural land

Arable land

4 million ha

800,000 ha

Share in ‘enterprises’ (corporate farms)

35%

20%

Share in household plots

20%

Share in dehkan farms

59%

60%

Estimated share in family dehkan farms (2/3 of land in dehkan farms)

39%

40%

Estimate for individual sector (household plots and family dehkan farms)

45%

60%

Estimate for corporate sector (enterprises and collective dehkan farms)

55%

40%

Total land

afford the official (and unofficial) costs of the application process are privileged to obtain such farms; hence, the process is an invitation to corruption.40 Agricultural output has increased quickly during the recovery since 1997 to recover and approach the 1988 peak of the late Soviet era, while agricultural land under cultivation has remained roughly constant (or has perhaps even declined slightly). Land productivity has thus increased substantially; agricultural labour has continued to increase, its recent productivity gain has been more moderate, and its level has remained below that of the late Soviet years. Despite its limitations land reform has played an important role in agricultural recovery. Table 4.2 presents an estimate of the 2006 breakdown of land among the different categories. Family dehkan farms together with household plots probably had 40–50 percent of agricultural land and a larger share (perhaps 55–65 percent) of arable land.41 Growth of crop production was driven by productivity increases from the household plots (Lermer and Sedik, 2008: 2) and a redistribution of land in dehkan farms. By 2006, with crop production 30 percent above its 1991 level, household plots accounted for 50 percent of total output, compared with 36 percent in dehkan farms and 14 percent in corporate farms. Since household plots held only 20 percent of the sown area, while corporate and dehkan farms controlled 80 percent, average land productivity on the former was four times greater than on the latter. The performance of household plots stands apart from the others also in that over 50 percent of growth in crop production on these farms can be attributed to rising productivity. If dehkan farms and agricultural enterprises had achieved the same level of land productivity as household farms in 2006, total agricultural production would have been 114 percent higher; if they had achieved even half that productivity level, total production would still have been 37 percent higher (ibid.: 4). It is noteworthy that dehkan farms were on average not doing better than the large farm enterprises. Part of the problem, no doubt, was that at least one third

Tajikistan

179

of the dehkan farms were not individual farms at all; it is plausible that these collective dehkan farms would perform at about the level of their corporate counterparts. Access to machinery and fertilizer was greatly curtailed after 1991, until at least 2005 for machinery, while fertilizer use seems to have stabilized somewhat earlier. This difficult setting further highlights the impressive achievement of household plots, generally at a greater

Figure 4.4: Productivity of Land by Farm Type, 1991–2006 (GAO per hectare of agricultural land, by type of farm, TJS per ha in constant 2003 prices, log scale)

Source: Lerman and Sedik, 2008.

Figure 4.5: Cropping Pattern in Household Plots and Dehkan Farms (DF), 2007–2008

Source: FAO 2007–2008 survey for household plots and dehkan farms; 2007 USAID survey for collective dehkan farms.

180

Social protection, growth and employment

disadvantage in getting access to machinery and inputs than the state-run corporate farms and the larger collective dehkan farms, although such plots typically do not require machinery and rely more on natural rather than mineral fertilizer. There is a striking contrast in the cropping decisions across these farm types, with the share of land dedicated to cotton drastically lower on individual/family dehkan farms (and on household plots) than on collective dehkan farms. In fact, farms were reorganized into collective dehkan farms predominantly to ensure the continuation of state-mandated cotton production on this land. Hukumat (local government) intervention is found to be quite prevalent for cotton growers and virtually non-existent on other farms.42 Although one of the main stated aims of the reform was to give the farmers freedom to decide what to grow,43 local officials told AAH that a government production plan was still in place for cotton. At the beginning of the year, each provincial hukumat gave each district hukumat a production target; despite the formal changes, in practice much continued to follow the Soviet style (Porteous, 2003: 9). What explained the discrepancy between government goals and small farmers’ behaviour with respect to cotton, a crop with such a large traditional comparative advantage that they might be expected to continue to grow it even without government mandates? Porteous (2003: 11) quotes one farmer as follows: “If I can take the profit from the cotton, I will grow it on all of my dehkan farm, all of my household plot, and all of my presidential land. But right now, I am not taking much profit, so I want to be free to grow my land in wheat and vegetables instead.”44 The history of cotton and its institutional structure since the Soveit era provides the background to this impasse. When, in the early 1990s, it became clear that the government could not on its own return the cotton sector to its Soviet level of operation, a consortium of international cotton buyers, including Paul Reinhart and Credit Suisse-First Boston, stepped in. Millions of dollars’ worth of loans were made against later cotton purchases. The funds were distributed to local ‘investors’ through the Tajik AgroInvest Bank, in which Reinhart then held a majority share. The investors used a system of futures contracts to channel the funds to farmers in the form of inputs (fertilizer, pesticide, fuel and seeds) and then purchase the cotton for resale to Reinhart. The contracts were based on an expected production quantity, which if exceeded translated into profits for the farmers or into debt if targets were not met; by 2009 this debt had risen to US$548 million, an amount almost equivalent to the total combined amount provided by the World Bank and the Asian Development Bank (ADB) in loans or grants to Tajikistan since they began operations in the country. Several factors explain why these debts rose to the point of crippling the cotton sector. First, Tajikistan has never been able to bring its cotton yields back to the level of the Soviet era; the predicted yield for 2012 is still under half that of the peak in the late 1980s.45 Since 1990, although the cropped area has remained relatively constant, output has suffered a major decline due to the huge drop in productivity. Another part of the story is institutional. The ‘investors’ who set up the futures companies developed enormous influence over the industry, including the market for inputs (by virtue of their loans to the cotton farmers) and the purchase and export of the cotton. With time this group further consolidated its grip by purchasing ginning mills as

Tajikistan

181

the government privatized them. In short, the cotton industry became a complete monopoly/ monopsony, using debt to reduce farmers’ market freedom and diverting the bulk of the rents from the farmers to the ‘investors’, the government and international cotton buyers. This pattern is reflected in the exorbitant prices of inputs to farmers (debtor farmers have no choice but to buy inputs from the investor they are indebted to) and farm gate cotton prices that are much lower than in other cotton-producing countries.46 Another symptom of the exploitation is the smuggling economy that has developed around cotton; there have been several instances of cotton being smuggled out of Tajikistan in its raw form to be sold at higher prices in neighbouring Kyrgyzstan or Kazakhstan. While the monopolistic nature of the cotton market and its adverse effects on the growth of the economy is widely acknowledged, its reform is politically difficult. The Government of Tajikistan, under the auspices of the World Bank, the ADB and the European Bank for Reconstruction and Development (EBRD), has begun to address the problem, but counter-pressures are strong. Keeping farm profits low contributes to government revenues, to investors’ incomes and to the personal incomes of some government officials, lending some plausibility to the view that the state’s regulatory apparatus is allied with the ‘investors’ in deliberately making it difficult for the farmers to extricate themselves from debt.47 As Rowe (2011: 201) puts it “the changes undertaken are the work of a government that does not want to change the economic, social or cultural status quo of the rural countryside.” The debt accumulated on these lands is one of the most cited reasons for why a larger number of farmers have not applied for their own private land. In May 2009, after several consultations and discussions with the donor community and the Government of Tajikistan, the President issued decree No. 633, which wrote off certain cotton debts to farmers and promised that by June 2010 land use rights would be strengthened by allowing farmers to engage in market transactions without state interference and by requiring court proceedings for the withdrawl of land use rights for state or public need when parties cannot agree. These reforms are very welcome; it now remains to be seen how thoroughly they will be implemented, particularly as uncertainty of maintaining tillage rights has been a major reason for the lack of investment and hence the stagnation of productivity. Most critical, however, is that the depressed cotton price to the producer due to the monopsonistic market must be tackled if productivity in the dehkan farm sector is to approach levels found in the household plots. Cotton remains important to the national economy. It accounts for 15 percent of Tajikistan’s export revenues and 39 percent of total state tax revenue, providing the government with an incentive to ensure that it continues to be grown as a major crop. But the structure of the cotton sector has been one of the biggest barriers to local investment in Tajikistan and has thereby contributed to migration in search of better opportunities in Russia. A majority of migrant labourers come from poor rural cotton-growing areas. Khatlon, the nation’s largest cotton-growing province (with 60 percent of the total), has by far the highest level of extreme poverty, defined as per capita income of under US$1.05 per day. The reform of the cotton market in Tajikistan is critical to maximizing the developmental payoff to land reform. More generally, the importance of reforming the agricultural sector is fundamental to achieving food security and equitable

182

Social protection, growth and employment

growth in the country. Freedom to decide which crops to sow, access to secure and transferable land use rights, and investments in rural infrastructure are all central to this goal. Currently, too many young Tajiks are deciding that it is pointless to apply for land ownership and that the best option is to leave farming and migrate to Russia to look for work there. Cotton, as a relatively labour-absorbing crop, has the potential to contribute importantly to pro-poor growth, provided the appropriate institutional (land reform, end of monopoly control) and technical (infrastructural investment, crop improvement) steps are taken. Whereas currently the workers toil with little benefit, accompanied by children who should be in school (International Crisis Group, 2005: 17), if cotton were produced on small family farms with solid technical back-up from the state and competitive (or otherwise not exploitative) input and product markets, incomes would be greatly increased, the crop’s contribution to employment would remain, but the exploitation of cotton workers and their need to have their children in the fields with them would end. Both the management of cotton and the societal benefits from it will be improved by continuing the shift towards small family farms as the basic unit of Tajik agriculture. Currently the government bias towards cotton production but against paying farmers a competitive price for it leads to much inefficiency. Rowe (2011: 193–194) notes in connection with the Hisor Valley, a more or less representative area in terms of agricultural land and setting, that workers on the better land in the central part of the valley where the primary crop is cotton have good access to water, but most farmers have had difficulty accessing it for household plots and other private land. They say that the bureaucracy is “too new, too confusing and too exhausting to effectively get anything done on land that does not grow cotton or other commercial crops”. While land reform has registered significant achievements in Tajikistan, it has not been able to achieve nearly as much of its potential as would be desirable and is essential for the welfare of the country. As Lermer and Sedik (2008: 4) put it, “…failure to follow through on land reform has prevented the government from attending to the longer run needs of agriculture, rural development and natural resource management… The failure to take the first and most basic step preserves an underperforming agriculture, keeping the rural population on the brink of food insecurity, agriculture particularly susceptible to natural disaster and government without adequate tax revenues from the sector.” In short, the failure of land reform has contributed to the vicious circle plaguing the country’s efforts to pull its population out of poverty. As in any country, the design of the agricultural system should reflect cultural patterns and history. Rowe (2011: 201) concludes that culturally most Tajik farm labourers are not ready for full privatization, given that production was a communal affair even before the Soviet period. As he proposes, Tajikistan might consider the dual system of Hungary (Meurs, 2001), under which private and communal property owners compete again each other, spurring the latter to improve to the level demonstrated by the former. The communal option would give greater security to the workers but on average less income.

Tajikistan

183

Small private enterprises outside agriculture About half of domestic employment in Tajikistan has in recent years been generated in small and medium-sized private firms, including those in agriculture. For much of the foreseeable future such firms and, increasingly over time, those outside agriculture (internationally referred to as micro, small and medium-sized enterprises—MSMEs) will have to carry the bulk of the burden for the domestic creation of decent jobs; agriculture’s role, though pivotal for the short to medium term, will gradually decline. In any case, small firms both in agriculture and elsewhere and both microenterprises (usually defined as family firms with up to about five workers) and SMEs (firms with five to 100 workers or so) will be needed to generate enough jobs, since even if the larger capital-intensive projects that are planned do come to fruition, they cannot realistically be expected to generate much employment. The MSME sector has been growing at least reasonably rapidly over the last 10 years or so,48 as have the incomes generated in it; productivity trends are not known. Such activities currently compete with work abroad (Abdulloev et al., 2011) but could eventually provide a more secure source of employment for many more people. Judging by the large number of self-employed people in the labour force, it appears that microenterprise provides the bulk of MSME employment, with SMEs still playing only a modest role.49 In qualitative terms Tajikistan’s SMEs have much in common with such firms in marketoriented developing countries. As expected, the main source of start-up capital is private or family funds (Olimov, 2007: 8), with loans from banks and other financial institutions playing a minor role50 because of the high cost of getting such loans and the high interest rates payable on them (24–30 percent per year). Small firms conduct many transactions in cash rather than through banks; about 80 percent of the enterprises did not use bank accounts in their business operations. Firms also prefer informal methods of dealing with state officials and other businesses because of the overcomplicated and burdensome formal procedures and the low quality of civil servants (ibid.: 9–10). Informal channels to solve disputes and other issues are more uniform and hence predicable than formal ones and require only a third as much time. Cash outlays are considered comparable between the two channels; informal payments (bribes) to solve disputes with state officials amount to 7.2 percent of sales or about a third of annual profits. Smaller firms have to make larger payments in relation to sales than do larger ones. The informal sector (or microenterprise) is a sort of school for individual entrepreneurship, providing opportunities in circumstances where operating a small or medium-sized business is prevented by administrative and other barriers. Employment in the informal sector can allow relatively ‘cheap’ labour to attain new skills while facilitating social, employment and occupational mobility. There appears to be considerable interest in starting new businesses. Thus according to the labour force survey of 2004, 70 percent of the total number of unemployed young people were actively looking for work, while the other 30 percent were trying to start their own business. A high level of sustained healthy growth of SMEs in a low-income transition economy such as Tajikistan’s will not be easy to achieve because of a combination of obstacles:

184

Social protection, growth and employment i) the limited entrepreneurial experience and skills, in part a legacy of the centrally planned economy; ii) the absence of a strong support system for new and evolving small firms, both in agriculture and elsewhere, and the presence of a not very competent and sometimes predatory state;51 iii) electricity shortages and a poor transportation network. Lack of electricity limits business activity, especially in winter, and the weak transportation system impedes the movement of goods between regions and abroad; iv) the monopoly control exercised by an in-group on some activities into which, ideally, small firms would enter in numbers; and v) the lack of some of the skills (other than entrepreneurship per se) needed on the part of employees and workers in such firms.52

Policies to foster growth, employment creation and social goals: the need to look beyond remittances Growth in remittance income cannot be expected to continue indefinitely, and its contribution to a rising national income will, therefore, likewise decline. Planning for that phase of the country’s development requires attention to which sectors have the potential to contribute the most to propoor growth, what knowledge and skills will be most important to success in those sectors and activities, and what social policies will be best able to complement the growth process thereby set in motion.

Key sectors The rapid relative shrinkage of both the manufacturing and the agricultural sectors does not bode well for the economy, given that Tajikistan is still at an early stage of its development.53 The manufacturing sector is particularly worrisome, given that its official share of value added (at current prices) slipped under 10 percent in the last couple of years and that one large aluminium smelter — a Soviet-era legacy established to take advantage of the low cost of hydroelectric energy—accounts for more than 40 percent of sectoral output. The agricultural sector accounted for 25 percent of the country’s GDP in 2008 but only 20 percent in 2011; even as the relative weight of this sector continues to decline for a myriad of reasons, it remains the largest employer, accounting for over half of total jobs. Much of the huge decline in labour productivity after independence occurred in the agricultural sector. Institutional failures, inefficient infrastructure and problems specific to the cotton sector all contributed to its decline. To achieve pro-poor growth and adequate employment creation, Tajikistan must be successful in a number of labour-intensive sectors, including cotton, other agricultural products,

Tajikistan

185

labour-intensive services and, possibly, some branches of manufacturing. The most obvious opportunities are in agriculture, where both cotton and other products could generate more and more productive jobs if the land reform process were restructured and accelerated. Since the interrelationships among the agricultural/rural sector, food security and emigration are key elements of the background scenario for the analysis of economic policy, social policy and labour markets in Tajikistan, any meaningful social protection policy must be designed with them in mind.54 Specific attention must be given to reforming the cotton sector, which has been shackled by poor governance, lack of access to credit, and a counterproductive system of land use rights. Overall agricultural productivity is also hindered by poor infrastructure maintenance, lack of market access, and deterioration of soil quality. Non-agricultural rural activity is severely constrained by inadequate access to energy and electricity. Policy recommendations need to be structured around the principle of achieving an adequate level of local food security by allowing crop diversification to modify the damaging aspects of cotton monoculture. Despite the severe challenges, pro-poor growth appears technically possible and would be fostered by the combination of: i) continuing strength of the labour-intensive (albeit seasonal) cotton sector, which will require policy to reverse the history of environmental degradation and to better organize the cotton chain such that farmers themselves receive more of the benefits. Cotton offers much promise, given Tajikistanâ&#x20AC;&#x2122;s excellent growing conditions and high-quality product. The cotton monoculture has been environmentally very destructive, and the benefits have been concentrated due to the structure of the cotton chain. Land reform to break up the large, inefficient farms that are a legacy of the Soviet economic structure is part of the answer. How much more labour-absorbing cotton production would be if carried out in small units is hard to judge at this point; ii) continuing reform of agriculture to get land into the smaller units that, according to experience both in Tajikistan and elsewhere, have the potential to be more efficient in their allocation and use of resources, more labour-intensive and also to focus more on food production for family consumption, with the effect of alleviating food insecurity, and also to foster more firms that create jobs; iii) developing a strong support system for non-agricultural MSMEs. Probably the most important policy step related to this sector is to avoid penalizing small firms, including microenterprise, either through unnecessary regulations and requirements or through more punitive measures. Despite some growth in the SME sector and the expectation that it will gradually become more formalized, the informal sector is likely to remain pivotal for employment generation for several decades; prohibitive measures against it would thus be counterproductive, possibly seriously so. Unfortunately some measures of this sort have been taken, such as closing small markets in cities where small producers come to sell their agricultural products. Attempts to discourage informal activity through high taxes and other impositions are likely to raise unemployment and push more people to emigrate. Income of workers in

186

Social protection, growth and employment the informal sector will be pushed down, and some people will inevitably resort to criminal activities. Restraining measures appear justified only in the most negative manifestations of informal employment, such as where it involves severe environmental damage, severe abuse of workers or criminal activities.

For SMEs also, it is important that policy not affect them negatively. One necessary step is to provide a simplified regulatory system, including easier procedures for registration of small businesses, tax cuts, and possibly exemption from taxes for a period after start-up. Measures to promote small businesses can contribute both to legalizing part of the informal sector and improving its accountability in terms of minimum social guarantees. On this proactive side, one useful step is to support the involvement of SMEs in the process of government procurement, which can be facilitated by reducing the batch size of public contracts to ensure that they are not squeezed out. Access of small firms might also be improved by requiring that all payments of state fees be channelled through the banking system to reduce vulnerability to bribes. Several minor initiatives have been undertaken to aid small business. In 2009, the State Agency of Social Protection, Employment and Migration assisted 1457 women and girls to start businesses by allocating to them soft loans totalling TJS2.944 million. In 2009, a number of unemployed women were involved in community work, and, through this, 14,811 found continuous employment, while 5541 were channelled into training courses for such female professions as seamstress, fabric cutter, gold embroiderer, weaver, clerk, cook, farmer etc.

Education and human capital A relatively good educational system was one of the positive legacies of the Soviet experience, but it has since suffered significant setbacks due to limited investments and the disruptive effects on attendance of the recent wave of migration. A recent UNICEF (2011) study shows that there are significant negative social impacts on children from migrant households, in particular on their psycho-social well-being and on educational outcomes. Especially in the case of abandoned households, the social costs of migration were found to be high, due to family disintegration and lack of parental care. The capacity of the public sector is limited by its very modest revenues and its limited administrative capacity. Low investment and spending on education translates into poor physical conditions of schools and low teacher salaries, which in turn contribute to instruction that is frequently of poor quality, lacking in appropriate teaching aids and updated tools for the transmission of knowledge, and of questionable relevance; it is also likely to negatively affect the perceived status of public education and to lead to a feeling among many teachers that they are being neglected by the government and are not respected by society (UNICEF, 2004: 8). Meanwhile, the dramatically lower incomes since the Soviet period have naturally constrained private expenditures on education.55 ADB studies show that the share of household expenditures going to education and health care is small, averaging no more than 3.8 percent. Before investing in education, households need to have a degree of confidence that it will improve employment opportunities. The lack of local opportunities even for the highly educated

Tajikistan

187

is believed to deter expenditures on schooling. Emigrants typically get jobs that do not require and/or reward much education. The effect of remittances on education is complicated. On the one hand, studies point to a strong positive relationship across households between expenditures on education and the level of remittances received, as well as between the level of educational attainment of the migrant and the remittances sent (Kroeger, 2010). On the other hand, there is a strong connection between receipt of remittances and absenteeism from school. Therefore, while remittance income can stimulate investment in human capital and relieve the financial constraints on the length of time spent in education, the associated loss of household labour may be increasing the pressure on remaining younger members to compensate by working when they could be in school. This is another of the dilemmas associated with emigration and remittances. The currently discouraging context for investment in education is worrisome, since it is plausable to assume that a better stock of human capital would contribute to future growth and that such growth would create the demand for better-educated workers. At present, however, a vicious circle may be at work, whereby the state of the economy discourages education, which in turn limits the paths the economy could take in future. SIDA (2007: 9–10) concludes, “Concerted action may …be essential today to achieve the desired results half a generation from now…”

Social policy during the transition and beyond Since independence, Tajikistan has experienced a strong deterioration of social-sector services (health, education and social protection, water, sanitation and other municipal services, and environmental protection) as well as in gender equality.56 The crisis has reduced by half the population’s access to quality social services and thereby contributed to a general decline in living standards. The state’s limited institutional and fiscal capacity has made it impossible to execute sound and comprehensive reforms for the social sectors, to strengthen their human capacity and to expand the logistical base for the social services. Existing staff in Tajikistan’s social-service institutions have little relevant experience. Training to remedy that deficit is an urgent need.57 With a shrinking public sector in relation to the size of the economy and a growing private one with low levels of social security coverage, Tajikistan faces the need both to reform the public-sector social security system and to design appropriate regulations for the private sector. The country is in the unusual situation where incomes are generally higher in the private sector, but, with pensions better in the public sector, some people have preferred public employment, especially when close to retirement. The weak business environment and the cumbersome administrative barriers impede the development of the private sector, which could help to provide some social services directly and strengthen the government’s fiscal capacity to provide guaranteed social services. Recent assessments have identified various groups of Tajikistanis particularly at risk. According to the National Social Investment Fund of Tajikistan (NSIFT) survey, the very poor people in Tajikistan include:

188

Social protection, growth and employment • families headed by women. Unlike men, who are more mobile and can more easily go to where work is available (including abroad), many female heads of families have to care for their children and can only access scarce job opportunities within their village; • families with young children, headed by disabled people. The situation in these families is even more difficult, because young children (under seven years old) cannot generate income. A disability pension, which heads of such families receive, is only enough to feed the family for two or three days, even if it is paid regularly, which is not always the case; • families with five or more children, especially if this family has young girls who do not work and cannot contribute to the family income. This situation arises especially in rural areas; and • lonely elderly people. Generally, children of these people do not have the resources to provide adequate help to their parents. The most vulnerable groups include: • families with a sick head of household. Illness of the breadwinner of a low-income family can lead to sudden poverty not only due to the loss of income but also the cost of treatment; • unemployed people who have lost a job and cannot find another. This group includes former employees of industrial enterprises, as well as young people approaching mid-career; • working poor people: families where a family member is working but whose income is too low to overcome poverty. Much of what he/she earns (80 percent) is spent on food. The rest is spent on clothing and other basic needs; and • farmers with small plots of cultivated land and small livestock or having no assets at all are mostly in poverty. Income derived by farmers from their small plots can often support their families for at most four or five months. The droughts afflicting the country since 2008 have particularly worsened the situation of these farmers.

Unfortunately, the challenge of reaching these groups in a satisfactory way is not matched by the programmes currently in place. As of 2009, the consolidated budget for social assistance (excluding social pensions) was tiny, at 0.2 percent of official GDP (World Bank, 2010: 1) and hence a considerably smaller share of correctly measured GDP or gross national product (GNP). Further, because this assistance is not well targeted, it exerts almost no downward influence on poverty rates or on income inequality; only 23 percent of the spending is likely to have reached the bottom quintile (although most leakage is to near-poor people), and a relatively small share of poor people receive any assistance at all — 20 percent in the case of the poorest quintile (ibid.: 10). The main social assistance programmes at this time were a subsidy for electricity and gas consumers and a conditional cash transfer (CCT) programme for families with children in school. The former suffered from the fact that many of the poorest families had no access to electricity or gas in the first place. The latter programme was set up in the early 2000s with World

Tajikistan

189

Bank assistance, with the aim of reaching the poorest 15 percent of families; having chidden in grades 1–9 was the criterion for eligibility. Unfortunately, its size is negligible—too small to have any likely effect on school attendance. Its administration is deficient as well; funds are transferred from the Ministry of Finance to the districts as part of a block grant, and the districts are not accountable for spending, so they can and do divert funds to other uses. The other main form of social assistance, managed by the state Agency for Social Insurance and Pensions, takes the form of pensions for elderly poor people who have not contributed to a pension fund (ibid.: 8). It amounts to 0.3 percent of official GDP; adding it to the other programmes brings the total spending up to 0.45 percent of official GDP, well below any other European or Central Asian country (ibid.: 6) and no doubt one of the lowest figures in the world. The World Bank (ibid.: 2) estimates that to eliminate extreme poverty would cost about 1.4 percent of official GDP, about four times what was being disbursed in 2009, even if the spending were targeted perfectly. In broad terms, therefore, what is needed is a major increase in spending and a good targeting system to replace the rather accidental processes currently in place.

Recent developments in the social policy area In 2007, a new programme was agreed on between the Government of Tajikistan and the European Union (EU), under which the EU provided policy support of €17.5 million (US$24.3 million) for 2007–2010. This constituted the final stage of the budget support programme to food security and the beginning of the programme of sectoral budget support, whose main objective was to assist the Government of Tajikistan to improve the design, implementation management and effectiveness of national social protection policies in line with its existing programmes, the National Development Strategy and Poverty Reduction Strategy Paper. The programme also aims to strengthen public financial management in the field of social protection. Currently (2010–2012) the EU has been implementing the Technical Assistance to Sector Policy Support Programme in the Social Protection Sector project, with a budget of €5 million and the objective of updating the social protection system. The main local partner is the Ministry of Labour and Social Protection. The programme has three components: • labour market policy, which aims to support a comprehensive analysis of the current situation in the labour market and labour migration for a labour market development strategy and implementing financial plans; • social payments and benefits policy, which aims to improve the system of social benefits and privileges in Tajikistan; and • social services development, which provides support for the development of standards, methodologies and practices in the provision of social care for people with disabilities and vulnerable groups, improving coordination and quality control of delivered services.

190

Social protection, growth and employment

Conclusion Tajikistan has for a little over a decade been experiencing an economic recovery from the severe crash that resulted from the break-up of the Soviet Union and the civil war that followed it. GDP per capita remains well below that of the later Soviet years; GNI per capita has performed considerably better due to emigration to and remittances from Russia. Poverty has fallen but remains the highest of the former SSRs of Central Asia. The most important policy for the country to achieve healthy pro-poor growth and continuing poverty reduction is to advance its family-farm strategy through further allocation of land into small units, reform of the marketing systems for agricultural inputs and outputs and development of a small-farm support system that can provide the infrastructure and the new varieties and breeds needed to raise productivity in this part of the economy. The other sectorspecific policy needed is support for non-agricultural MSMEs. In part this entails simply making the bureacracy less of an impediment than it has sometimes been; the other part is to develop a positively supportive set of programmes and policies. Development of the countryâ&#x20AC;&#x2122;s hydroelectric potential is a high priority both as an input to industry and as key to the provision of electricity and water to the population. But since they create few jobs, such capital-intensive activities must be balanced by labour-intensive ones (such as family agriculture and MSMEs); otherwise, any growth that results will neither create employment nor be pro-poor. The health and educational systems and the infrasructure behind them are also significant challenges, given the deterioration that has occurred. Finally, a social protection system consistent with the current state and trends of the economy will need attention for the foreseeable future. It will be important that the pursuit of such a system does not come at the expense of successful evolution, including eventual formalization, of smaller private firms through onerous regulations. Among the high-priority elements are contributions to food security (above and beyond the development of a strong family-farm system), improvements in primary health care, and programmes to assist the poorest and most vulnerable members of society. The multiple crises that have recently afflicted Tajikistan help both to understand the challenges confronted and also to provide an opportunity to develop policy that is more than a simple reaction to them. Improving food security requires going beyond short-term responses, to protect and promote peopleâ&#x20AC;&#x2122;s livelihoods in the longer term. Issues such as negotiating access to markets or dealing with land rights and occupation are long-term issues, in contrast to much of the humanitarian work necessitated by short-term crises.

Tajikistan

191

Annex 4.1: The informal economy and mismeasurement of aggregate economic variables Although the enormous and changing underground economy (including legal but informal activities as well as illegal ones) makes it impossible to asecertain with precision what has been happening to true GDP or GNI, several serious attempts have been made. Olimov (2007: 8) estimates that for 2005 the shadow economy was 61 percent as large as officially reported GDP. The three components he distinguished were home production of food products (14.7 percent), wages paid in kind and barter exchanges (13.2 percent) and understatement of salaries (35 percent), with the related underpayment of taxes amounting to 33 percent. These economic activities uncaptured or understated in the national accounts do not include, or do not fully include, informal enterprises or illegal activities. It is thus possible that the total size of the underground or unreported economy is equal to or larger than the official one. A different type of evidence on the degree of underestimation of economic aggregates is presented by Abdulloev et al. (2011) on the basis of the World Bank’s 2007 Living Standards Measurement Survey; they point to huge underreporting of household incomes –which are only 41 percent of reported expenditure (which in turn is likely to be somewhat underestimated). This enormous gap is broadly consistent with the conclusion of Olimov, allowing additionally for the components of underreporting that he did not try to estimate. Abdulloev et al. interpret the difference between household income and household expenditure as informal income. In the absence of huge under-measurement of other forms of income, the numbers do point strongly to the importance of informal income; even if other forms of income were understated by, say, 50 percent, there would still be 36 percent of expenditure to be accounted for and presumably explicable by informal income. As noted in the text, national income and GDP may also be significantly underestimated due to failure to capture drug-related income; the same could be true for household income, where drug-based revenues are unlikely to be reported. It is important, however, to note that there are also upward biases in the official GDP figures, in particular those related to its failure to allow for environmental degradation. Adjusted net savings, which take account of the depreciation of produced capital, investments in human capital and depletion of natural capital offer a truer depiction of the level of saving in an economy. A World Bank estimate put this figure at 2.8 percent of GDP saving in an economy (2011), although this calculation omits two of Tajikistan’s most eroded resources: water and soil.

Notes 1 The Food and Agriculture Organization of the United Nations (FAO) estimated that, because of the food price inflation in 2007 and 2008, an additional 115 million people were pushed into chronic hunger (2009: 6).

192

Social protection, growth and employment

2 Figures appear to vary, perhaps because of differing assumptions made about the number of Tajikistanis abroad. 3 For more details, see Annex 4.1. 4 Thus if, for example, remittances were zero in 2000 and 30–50 percent of GDP in 2010, then GNI growth over 2002–2010 was about 3–5 percent per year above the correctly measured growth of GDP, which in turn may have exceeded the officially reported rate for the reasons just mentioned. 5 Informal-sector activities often have high output–capital ratios, so failure to include this sector in the official figures might imply a smaller understatement of investment than of output, and hence suggest that the official investment ratio from the official figures overstates the true one. But the pattern might be different for illegal activities. 6 Olimov and Bosc (2003: 36) distinguish hired workers (working under some form of formal or informal contract), construction workers, shuttle workers who buy goods in one region and sell them in another, and agricultural workers. 7 ILO (2010: 1) notes that, according to the Ministry of Labour and Social Protection of Population and the Migration Service of the Ministry of Interior Affairs, the total number of migrants in 2008 was 720,000, of whom 93 percent worked in the Russian Federation and 5 percent in Kazakhstan. Mughal (2007: 111), extrapolating from a household survey in the region of Khatlon, put the figure at just 371,000 for 2005. 8 For a comparison among sources, see Jones et al. (2007: 21). As might be expected, the highest figure comes from a survey of migrant households, which would presumably capture all forms of remittances. 9 Three figures are worth bearing in mind. First, if remittances are 45.4 percent of GDP, and disregarding other differences between GDP and GNI, the latter would be 45.4 percent above the former. In fact, however, national income also includes earnings of emigrants that they do not remit but rather spend in the country of work. This adds another substantial amount to national income—say, another 50 percent of the value of the remittances—raising GNI to about 170 percent of GDP. Finally, however, the impact of emigration on GNI is less than this by the amount of income the emigrants would have earned had they stayed in the country. If income abroad is three times that which would have been earned at home, then GNI would in fact be about 147 percent higher than if no migration had occurred. Finally, however, it must be noted that both the official GDP figures and any GNI figure based on them are likely to be a serious underestimate through failure to include informal and illegal (e.g. drug-related) income, which between them probably amount to 60–100 percent of official GDP. In that case remittances, at 45 percent of official GDP would be 22.5–28 percent of true GDP and 17–20 percent of GNI. 10 If these families had been unable to substitute other forms of income for the remittances, extreme poverty would have leapt to 77.8 percent. Since any such substituion would have been at most partial, there is no doubt that poverty would have been much higher than in fact it was. Economic theory suggests that the short- to medium-term impacts on poverty result from three phenomena: i) tightening of the local labour market, thus pulling wages/earnings up; ii) upward impact of the remittances on the incomes of those left behind and the emigrants themselves; and possibly iii) a selectivity according to initial level of poverty. The third factor could work either way. If enough skilled workers or entrepreneurs migrate, this could have the effect of lowering the demand for labour in the unskilled categories where poorer people would be found. In Tajikistan, it is not clear whether the emigrants come from ex ante lower-income families or not; ex post they have a higher average income. In the

Tajikistan

193

longer term, the impact on poverty also reflects the impact of the emigration/remittance process on growth, to which we turn below. 11 Note that the available figures on income inequality might be expected to understate that variable, perhaps by a large amount given the prominence of monopoly power and of the drug trade, both important sources of income concentration. Some surveys do not appear to have included remittances as part of income (Korobkov and Paley, 2005: 139), leading to very high reported Gini coefficients. Thus Falkingham and Klytchnikova (2006: 13), while reporting a per capita expenditure Gini coefficient of 0.36 for the country as a whole, show a coefficient of 0.63 for per capita income. As per the normal relationship between these two indicators, one would expect the Gini coefficient of per capita income to be in the range 0.4–0.45. 12 Available sources vary in their apparent coverage of income sources. Falkingham and Klytchnikova (2006: 13) report a Gini coefficient of per capita household expenditure of 0.36 in 2003, with the true figure presumably higher due to measurement error. The corresponding Gini coefficient of income would probably be somewhat above 0.4; allowing for the usual downward bias due to measurement error would push it somewhat above that level. 13 Jha et al. (2008) note that as of 2004–05 more than half of households observed to be non-poor were still vulnerable to poverty. 14 Such tactics help to maintain Russian influence in a country that has repeatedly challenged Russia by openly supporting Iran, ceding land along its eastern border to China and aiding the USA and its allies with the war in Afghanistan. 15 For reasons not clear to us, the figures used by these authors suggest a much lower ratio of remittances to total incomes than is commonly accepted in the literature. 16 Such concerns have been evaluated by analysts looking at migration and remittances in other countries. Adams (2011: abstract) concludes that “while international remittances generally have a positive impact on poverty and health in the developing world, remittances can have negative effects on labour supply, education and economic growth.” 17 www.asiaplus.tj/news/198/66314.html . 18 The majority of this group, or 35 percent of the total, are long-term migrants (away for over two years with only brief trips back), while 16 percent have settled down in the other country, making no home visits over two years; some of these emigrants are unlikely to return or to maintain their current level of remittances (ILO, 2010: 16). 19 As an economy with a high level of temporary emigration, Tajikistan is somewhat like Puerto Rico at an earlier time; that country achieved very fast growth by inducing investment, much of it from the USA (Wolff and Baumol, 1996). Jamaica, on the other hand, did not reap such notable benefits. 20 Some of the funds arriving as remittances have indirect effects—for example, those that wind up in the hands of financial establishments and government. It is naturally harder to trace those effects. It is also possible that local spending for consumer goods stimulates local production and employment. 21 Comparison with the experiences of the countries analysed by Vaaler (2011) could prove helpful to explore this relationship further.

194

Social protection, growth and employment

22 Although some forms of unregistered employment existed in Tajikistan during that period, they did not account for a large share of total employment. 23 Although labour market data are hard to interpret in Tajikistan, when pieced together they probably give a rough idea of recent trends. Official figures are less meaningful than before, given the absence of information on the informal sector or a large part of it; this means also that the official data on the sectoral composition of employment are unreliable. As Babajanov and Khakimov (2010: 26–27) note, “Even with the clarified concepts and improved approaches and techniques, many of the statistics will be flawed due to technical reasons. (…) Recently, household surveys are increasingly being used as a source of data, but this cannot fill the statistics gaps and address in full the mentioned shortcomings.” 24 As in other countries, the distinction between ‘formal’ and ‘informal’ involves not so much a dichotomy as a spectrum reflecting degrees of adherence to the various types of regulations on firms’ behaviour. 25 Thus the household survey used by Olimov (2007: 9) gives a figure of 46 percent for 2006. 26 Formal unemployment (registered at an employment office) was at 52,000 (or a little over 1 percent of the working-age population and 2.4 percent of the labour force) in 2007. These figures by themselves tell us little about the state of the labour market. 27 They report that during the rapid downsizing of the public sector in Bulgaria the informal sector expanded more rapidly than did the formal private one, in a context of complex and gender-differentiated links between these two. Government policy had the effect of making the cost of employment in the formal private sector higher than it might have been. 28 There are frequent cases of workers closer to retirement age trying to move from the private to the public sector to receive a higher government pension. 29 Table 4.1 provides another window into the sources of work-based income. 30 The ‘eastern mentality’ and certain accepted national traditions fit well with informal employment and thus constitute another advantage compared to working in the formal sector (Babajanov and Kakamov, 2010: 67). 31 These numbers are not comparable to those of Table 4.1, but they imply that each of the items in the table would be smaller by perhaps a quarter if the base were total income rather than cash income. 32 This suggests either that total household income is about equal to GDP or that the author is assuming it is. 33 Olimov (2007: 60), based on the Tajikistan Living Standards Survey of 2003 or 2007. 34 Thus women’s average earnings vary between less than half of men’s in the health sector to about three quarters in education, science, culture and art. 35 This paragraph is based on Babajanov and Khakimov (2010: 60–61). 36 IFPRI (Msangi and Rosegrant, 2008: 9) estimated that 30 percent of the rise in cereal prices in 2008 was attributable to the effect of demand for biofuels. 37 Data from 2000 indicate that about 90,000 ha of irrigated land (out of 720,000 ha irrigable) were in an ‘unsatisfactory’ condition— about half due to water-logging, a quarter to salinity, and the remaining quarter to both (Ministry of Irrigation and Water Management et al., 2006). A good review of the tight links between Tajikistan’s rural poverty and the condition of the land for farming is Swarup (2009).

Tajikistan

195

38 A recent disaster in the town of Kulob, where a mudslide caused numerous casualties, could possibly have been avoided had the community’s energy needs been met through regular electricity supply. 39 Tajikistan was the third largest cotton producer in the USSR during the 1980s (after Uzbekistan and Turkmenistan), producing 10 percent of raw cotton and more than 25 percent of fine-fibre cotton. 40 In their interviews with AAH, several jamoat chiefs (who owned their own independent farms) mentioned certain well-to-do individuals who owned several independent farms each, all registered under the names of different family members. 41 If family dehkan farms are assumed to have two thirds of the land in all dehkan farms, the figures are those shown in Table 4.2. These figures are not known with precision, since the database from which they come was partial (Lermer and Sedik, 2008: 2). 42 Among dehkan farms growing cotton, only 14 percent have freedom of decision, whereas in 56 percent the decision is made by the manager and in 28 percent by the hukumat directly. This is in striking contrast to the decision-making process in other dehkan farms, where 60 percent make the decisions themselves and the hukumat only intervenes in 5 percent of the cases (Zvi and Sedik, 2008: 51). 43 Article 5 of the law ‘On Dehkan Farms’ states, “Interference in the management of the activity of the dehkan farm from the side of state agencies and officials is not permitted...” 44 Other reports state that farmers growing cotton can earn far less than those able to grow other crops (International Crisis Group, 2005: 9). 45 See www.indexmundi.com/agriculture . 46 AAH’s study of the financing of the cotton sector reveals that in 2003 Paul Reinhart paid in the vicinity of US$400 per ton for raw cotton. Meanwhile, according to the investors, they provided somewhere between US$250 and US$350 per ha in terms of inputs and salaries to the farms (these numbers are likely to be inflated, as one of the most cited methods by which these investors have profited is by charging up to double the local market price for inputs), which yielded them up to 2 tons of raw cotton. This means that, by a generous estimate, the investors paid US$150 per ton of raw cotton that they were selling for US$400 as an export (Porteous, 2003: 14–15). 47 Clearly foot-dragging by and in the interest of influential people has played a central role. As Nissen (2004, quoted by Lipton, 2009: 225) puts it, “…as the process was getting under way…influential and educated individuals saw the personal gains to be made and took control of large tracts of land… Some were heads of local authorities [and] have been keeping farmers in the dark about their rights.” 48 Olimov (2007: 8) notes that the 500 enterprises surveyed for a UNDP study reported fast average growth of capital assets, sales, exports and productivity levels between 2003 and 2006 (these variables rising by 19–28 percent for an average of 6–9 percent per year). 49 Babajanov and Khakimov (2010: 102) report that “in Tajikistan there are about 1.4 small businesses per 1000 people, whereas in the Russian Federation there are six, and in the European Union at least 30.” The definition of a small business used in this comparison is not provided. 50 Olimov’s sample of 500 firms was weighted towards larger ones, with only 171 having 10 or fewer workers (34.2 percent) and with a median size of around 20–25 workers (Olimov, 2007: 26); it was thus representative of SMEs rather than of microenterprises. Sources of start-up capital among these firms included 146 loans from a financial institution (ibid.), and some of these were probably multiple. If,

196

Social protection, growth and employment say, 100â&#x20AC;&#x201C;125 firms got such loans, and they were the largest, that implies that very few firms currently with fewer than 100 workers did so. (On average, however, the firms would be smaller at the time of the loan, since they have grown in the interim.)

51 For the (mainly) SMEs in the sample analysed by Olimov (2007: 38), the share of working time spent on disputes with civil servants was over 5 percent for all but 23 percent of them, and over 15 percent for 31.4 percent of them. 52 SMEs cite the lack of finance and high interest rates as most frequent barriers to expansion, along with the low quality of the civil service, the shortage of electricity, weak transportation and onerous licensing requirements. Inadequate worker skills were not considered so serious a problem. Monopolistic practices and crime received medium billing (Olimov, 2007: 37). Unlike some other countries, labour market regulations do not seem to constitute a large barrier to the expansion of these firms, which report that workers often prefer to be hired on a part-time basis (ibid.: 40). 53 Based on the not very reliable national accounts figures, the bulk of the decline in the current price shares of these two sectors in GDP has come from falling relative prices. 54 The country has undertaken to implement two poverty reduction strategies designed under the tutelage of international financial institutions, and both have listed the agricultural sector as one of considerable focus. 55 Since economic growth was rapid between 1998 and 2011 (at an average of 7.7 percent, according to the World Bank Database, with some years even showing an annual GDP growth as high as 22 percent (2007)), one might guess that the willingness and capacity to invest in education would have risen as well but that, like income levels, it might not yet be at pre-crisis levels. 56 Under the Soviet system the minimal pension was 70 percent of the minimum wage, which was higher than it currently is in relation to a minimum acceptable consumer basket. At that time the system of social compensation reached a large proportion of the population (Babajanov and Khakimov, 2010:Â 83). 57 Improving Tajikistanâ&#x20AC;&#x2122;s social services infrastructure will, among other things, require the design, provision and coordination of computer databases and networks, the organization of seminars on various social issues, conferences at which social-service personnel can exchange ideas on solving existing problems and so on.

Tajikistan

197

References Abdulloev, I., I.N. Gang and J. Landon-Lane, ‘Migration as a Substitute for Informal Activities: Evidence from Tajikistan’, IZA Discussion Paper, No. 6236, IZA, Bonn, 2011. Adams, R.H., ‘Evaluating the Economic Impact of International Remittances on Developing Countries Using Household Surveys: A Literature Review’, Journal of Development Studies, 2011, Vol. 47, No.6: 809–828. Adams, R.H. and J. Page, ‘Do International Migration and Remittances Reduce Poverty in Developing Countries’, World Development, 2005, Vol. 33, No.10: 1645–1669. Baumol, W.J. and E.N. Wolff, ‘Catching up in the post-war period: Puerto Rico as the fifth “Tiger”?’, World Development, 1996, Vol. 24, No. 5. Brown, R., S. Olimova and M. Boboev, ‘Country Report on Remittances of International Migrants in Tajikistan’, A Study on International Migrants’ Remittances in Central Asia and South Caucasus, Asian Development Bank, Manila, 2008. Dimova, R., I.N. Gang and J. Landon-Lane, ‘The Informal Sector During Crisis and Transition’ in B. Guha Khasnobis and R. Kanbur (eds), Informal Labour Markets and Development, Palgrave-McMillan Press, 2006: 88–108. Falkingham, J. and I. Klytchnikova, ‘The Profile of Poverty in Tajikistan: An Update 1999 to 2003’, Applications & Policy Working Paper, No. A06/02, Southampton Statistical Sciences Research Institute, Southampton, 2006. Food and Agriculture Organization of the United Nations (FAO), The State of Food Insecurity in the world: Addressing food insecurity in protracted crises, FAO, Rome, 2010. Food and Agriculture Organization of the United Nations (FAO), The State of Agricultural Commodity Markets 2009: High Food Prices and the Food Crisis—Experiences and Lessons Learned, FAO, Rome, 2009. Intergovernmental Panel on Climate Change (IPCC), Climate Change 2007 — Impacts, Adaptation and Vulnerability: Contribution of Working Group II to the Fourth Assessment Report of the IPCC, Cambridge University Press, Cambridge, 2007. International Labour Organization (ILO), Migrant Remittances to Tajikistan: The Potential for Savings, Economic Investing and Existing Financial Products to Attract Remittances, ILO, Moscow, 2010. Jha, R., T. Dang and Y. Tashrifov, ‘Economic Vulnerability and Poverty in Tajikistan’, Economics RSPAS, Departmental Working Papers, Australian National University, Canberra, 2008.

198

Social protection, growth and employment

Jones, L., R. Black and R. Skeldon, ‘Migration and Poverty Reduction in Tajikistan’, IDS Working Paper, No. C11, University of Sussex, Institute for Development Studies, Brighton, 2007. Korobkov, A. and L., ‘The Socio-economic Impact of Migrant Remittances in the CIS’, International Migration Trends, 2005, Vol.15: 130–148. Kroeger, A., ‘Do Remittances and Migration promote Children’s Schooling in Tajikistan’, Preliminary Version, DIW Berlin, Berlin, 2010. Lerman, Z. and D. Sedik, ‘The Economic Effects of Land Reform in Tajikistan’, Policy Studies in Rural Transition, No. 2008-1, FAO Regional Office for Europe and Central Asia, Budapest, 2008. Lipton, M., Land Reform in Developing Countries; property rights and property wrongs, Routledge, London, 2009. Msangi, S. and M. Rosegrant, World agriculture in a dynamically changing environment: IFPRI’s long-term outlook for food and agriculture under additional demand and constraints, International Food Policy Research Institute (IFPRI), Washington, DC, 2008. Meurs, M., The evolution of agrarian institutions; a comparative study of post-socialist Hungary and Bulgaria, University of Michigan Press, Ann Arbor, MI, 2001. Ministry of Irrigation and Water Management of the Republic of Tajikistan, UNDP Office in the Republic of Tajikistan and Executive Committee of the International Fund for Saving the Aral Sea, Water Sector Development Strategy In Tajikistan, UNDP, Dushanbe, 2006. Mughal, A.-G., Migration, Remittances, and Living Standards in Tajikistan: A Report Based on Khatlon Remittances and Living Standards Measurement Survey (KLSS 2005), International Organization for Migration, Dushanbe, 2007. Naylor, R.L. and W.P. Falcon, ‘Food Security in an Era of Economic Volatility’, Population and Development Review, 2010, 36(4): 693–723. Nissen, S., ‘Tajikistan’s land reform programme: sowing the seeds of a brighter future’, OSCE Magazine, July, 2004. Olimov, J., Informal Economy in Tajikistan, UNDP and WAPK Research Centre ShARQ, Dushanbe, 2007. Olimov, J. and I. Bosc, Labour Migration from Tajikistan, International Organization for Migration, Dushanbe, 2003.

Tajikistan

199

Porteous, O., Land Reform in Tajikistan: From the Capital to the Cotton Fields, Action Against Hunger, Dushanbe, 2003. Rowe, W.C., ‘Agrarian adaptations in Tajikistan: land reform, water and law’, Central Asian Survey, 2011, Vol. 29, No. 2: 189–204. Save the Children, Implication of Labour Migration on Children in Tajikistan, Save the Children, London, 2008. Swedish International Development Agency (SIDA), Making Pro-Poor Growth in Tajikistan: An Integrated Economic Analysis Approach, SIDA, Stockholm, 2007. Son, H.H., ‘Evaluating Social Protection Programs in Tajikistan’, ADB Economics Working Paper Series, No.274, Asian Development Bank, Manila, 2011. Swarup, A., Reaching Tipping Point? Climate Change and Poverty in Tajikistan, Oxfam International, Dushanbe, 2009. Timmer, P.C., A World Without Agriculture: The Structural Transformation in Historical Perspective, The AEI Press, Washington, DC, 2009. UNICEF, Qualitative Survey of issue in girl’s education in Tajikistan, Ministry of Education and UNICEF, Dushanbe, 2004. UNICEF, Impact of Labour Migration on “Children left behind” in Tajikistan, UNICEF, Dushanbe, 2011. USAID/World Bank, ‘Knowledge, Attitudes and Practices of Farm Workers and Farmers Concerning Land Use Rights and Farmland Restructuring’, Tajikistan: Baseline Survey and Qualitative Study Findings, USAID, World Bank and the Government of Tajikistan, 2008. Vaaler, P.M., ‘Immigrant remittances and the venture investment environment of developing countries’, Journal of International Business Studies, 2011, Vol. 42: 1121–1149. World Bank, The Changing Wealth of Nations: Measuring Sustainable Development in the New Millennium, World Bank, Washington, DC, 2011. Zvi L. and D. Sedik, Policy Studies on Rural Transition No. 2008-1. The Economic Effects of Land Reform in Tajikistan, FAO Regional Office for Europe and Central Asia, Budapest, 2008.

5 Labour, Income and Social Programmes in Contemporary Mexico Carlos Salas Introduction Mexico is currently the second largest economy in the Latin America and Caribbean (LAC) region and the 14th worldwide (World Bank, 2012). Over the last three decades, however, the economy’s performance has been sluggish, with average annual growth of only 2.2 percent (1981–2011), 1.4 percent in terms of per capita gross domestic product (GDP), and no net gain at all in terms of output per worker. After a period of structural adjustment Mexico became an attractive destination for foreign capital, leading to a massive inflow. Since the 1994 crisis the country’s macroeconomic management has improved, judged by traditional criteria, as inflation has been reduced and interest rates brought to record lows, but growth has remained disappointing, at an average rate of 2.5 percent. The economy has recently rebounded from the brief but deep impact of the 2008–09 global economic and financial crisis (see Table 5.1). Large and slowly declining gaps remain between rich and poor people (see Table 5.2), particularly between rural and urban areas. Despite slow economic progress, a middle class has been growing since the late 1990s (Lora and Fajardo, 2011). But such a slow growth rate has impeded the rapid creation of jobs that would both pay a reasonable wage and provide workers with adequate job protection. Levels of poverty were falling prior to the onset of the crisis of 2007–2009 but then rose again through at least 2010, a trend accompanied by an alarming growth of poorly paid jobs and the failure of open unemployment to return to pre-crisis levels,1

202

Social protection, growth and employment

Table 5.1: Macroeconomic Data on Mexico, 1980–2011 Series Name

1980

1985

1990

1995

2000

2005

GDP (constant LCU) in trillion

4.40

4.84

5.26

5.68

7.41

8.11

MA5 GDP (constant LCU)

5.31

0.83

3.41

1.92

2.20

3.12

63,996

63,601

62,444

61,542

74,085

76,200

3.38

- 0.78

1.82

0.53

1.05

2.10

14.10

12.45

13.61

12.38

17.77

15.51

2006

2007

2008

2009

2010

2011

8.53

8.81

8.92

8.36

8.82

9.17

GDP per capita (constant LCU) MA5 GDP per capita (constant LCU) Gross Fixed Capital Formation, Private sector (% of GDP) Series Name GDP (constant LCU) in trillion MA5 GDP (constant LCU) GDP per capita (constant LCU) MA5 GDP per capita (constant LCU) Gross Fixed Capital Formation, Private sector (% of GDP)

2.55

0.60

0.67

0.80

79,120

80,664

80,586

74,609

777,62

79,859

1.51

- 0.42

- 0.35

- 0.20

16.55

16.50

16.52

15.21

14.24

20.48

Source: Data taken from the World Bank Database. The Moving Averages (MAs) are based on the author’s own calculations and encompass a timeframe of five years. The five-year MA for 2005 is the average over 2003–2007 inclusive.

with new employment opportunities being created mostly in establishments of five or fewer workers, in which working conditions and levels of pay are relatively precarious. In addition to reducing income inequality, government challenges include upgrading infrastructure and modernizing the tax system. A recent reform of labour law lowered the cost of firing workers, allowed for part-time contracts per hour and regulated outsourcing processes. Nevertheless, the new law makes it more difficult for workers to achieve the needed time for retirement payments, and leaves the supervision of labour conditions with firms outsourced by the contracting firms themselves. Various social programmes aiming to combat poverty, promote the creation of jobs or provide training have provided some help in alleviating the difficult working conditions and inadequate incomes of many Mexicans. But there are questions about the scope and relevance of these programmes in relation to the challenges to be addressed. With roughly 1 million Mexicans entering the labour force each year (Zepeda et al., 2009: 10), the generation of new jobs plays a crucial role. But the economy has provided far too few new jobs with social protection and good wages to absorb those displaced from agriculture and those newly entering the workforce (ibid.). This chapter focuses on the main patterns of employment in Mexico between 1995 and 2010, emphasizing the importance of job creation in smaller enterprises and how employment characteristics—including access to social protection—vary by size and type of business. The paucity of information on the reach and the impact of social programmes at the national level has led to uncertainty as to how to characterize the relative income levels and the overall labour

Mexico

203

Table 5.2: Poverty Data for Mexico, 1984â&#x20AC;&#x201C;2010 Indicator Gini index

1984

1989

46.3

1992

1994

1996

1998

2000

51.1

51.9

48.5

49.0

51.9

Income share held by: fourth 20%

21.4

26.1

19.8

19.3

20.4

20.7

19.4

highest 10%

35.1

16.9

40.3

41.5

38.0

37.7

41.4

highest 20%

51.7

32.1

56.0

56.9

53.7

53.8

56.6

lowest 10%

1.87

1.61

1.63

1.67

1.74

1.55

lowest 20%

4.67

6.37

4.09

4.13

4.43

4.09

3.98

second 20%

8.68

14.8

7.71

7.58

8.35

8.21

7.72

third 20%

13.6

20.7

12.4

12.1

13.1

12.3

Poverty gap at $1.25 a day (PPP) (%)

2.95

4.32

0.79

0.47

1.88

2.22

1.47

Poverty gap at $2 a day (PPP) (%)

9.64

4.65

4.25

3.56

6.42

6.78

4.74

Poverty headcount ratio at $1.25 a day (PPP) (% of population)

12.8

3.98

4.83

3.56

7.86

8.62

5.53

Poverty headcount ratio at $2 a day (PPP) (% of population)

28.5

6.55

15.2

14.0

20.1

15.1

Poverty headcount ratio at national poverty line (% of population)

53.0

53.5

53.1

52.4

69.0

63.7

53.6

Poverty headcount ratio at rural poverty line (% of rural population)

66.5

69.3

80.7

75.9

69.2

Poverty headcount ratio at urban poverty line (% of urban population)

44.3

41.2

61.5

55.9

43.7

2005

2010

Note: PPP = purchasing power parity. Source: Data taken from the World Bank Database.

Indicator

2002

2004

2006

2008

49.7

46.1

48.1

48.3

fourth 20%

19.9

21.0

20.0

19.9

highest 10%

39.4

35.6

38.3

38.7

highest 20%

54.8

51.2

53.6

53.7

lowest 10%

1.75

1.68

1.96

1.81

lowest 20%

4.39

4.55

4.80

4.73

second 20%

8.13

9.16

8.51

8.65

third 20%

Gini index Income share held by:

12.7

14.1

13.0

13.1

Poverty gap at $1.25 a day (PPP) (%)

0.77

0.13

0.34

Poverty gap at $2 a day (PPP) (%)

3.68

1.80

1.00

1.29

Poverty headcount ratio at $1.25 a day (PPP) (% of population)

3.90

1.60

0.68

1.15

204

Social protection, growth and employment 2002

2004

Poverty headcount ratio at $2 a day (PPP) (% of population)

Indicator

13.5

7.6

Poverty headcount ratio at national poverty line (% of population)

50.0

47.2

Poverty headcount ratio at rural poverty line (% of rural population)

64.3

Poverty headcount ratio at urban poverty line (% of urban population)

41.2

2005

2006

2008

4.94

5.19

2010

47.0

42.7

47.7

51.3

57.4

61.8

54.1

60.3

60.8

41.1

38.3

35.6

40.1

45.5

situation of potential target populations of these programmes. The chapter aims to throw new light on who the beneficiaries are of a set of public policies that create both employment and income. Mexico and countries that broadly share its level of development and labour market features, especially its high shares of very small economic units (including self-employment) and the low levels of social protection for waged workers (the so-called ‘informality’), constitute the laboratories within which several important challenges to policy are being played out. These include: • the impact of reforms aimed at making the labour market more flexible on the growth of jobs that enjoy social protection; • the impact on poverty levels of active labour market policies; • the direct social benefits of widening the coverage of social protection programmes; and • the success of programmes designed to raise the productivity of the very small units. Santiago Levy (2008) put forward the hypothesis that widening the coverage of social protection programmes to self-employed people would encourage low levels of social protection in paid employment. This thesis remains under debate, with studies such as by Knaul et al. (2012) contesting that view. Juarez pushed forward the argument that for low-income female workers the existence of a free health programme implies a lower level of women in jobs with access to social protection. Nevertheless, no proof of causality can be found in her research. As the evidence needed to prove the ‘good intentions, bad outcomes’ thesis2 is not yet forthcoming, there is a need to improve case studies and experimentation to sort out the effects of existing programmes and enhance the design and identify better versions of the first-round programmes. Economic theory does not provide many hints unless a good number of empirical parameters can be estimated with a reasonable degree of accuracy. The chapter is divided into six sections. The first gives a brief overview of the macroeconomic context in Mexico and highlights some of the factors that account for the pattern of slow growth that has been observed, in particular over the past decade. The second section

Mexico

205

examines recent developments relating to labour and unemployment; the third addresses the sectoral structure of employment by size of enterprise, social and labour benefits, and income, over 1995–2010; and the fourth focuses on the movement of workers between employment, unemployment and inactivity and between units of various sizes. The fifth section considers the behaviour of micro-units, highlighting the distinction among three subsectors: one that is business-oriented, another survival-oriented, and a third ‘transition’ category, then characterizing the latter two low-level income groups in terms of their access to income and social programmes. The sixth section considers various productive programmes and social policies that have an impact on employment. The study concludes with a synthesis of its major findings, which postulates a number of consequences for the formulation and implementation of social policies.

The macroeconomic environment After the rapid economic growth of the decades leading up to the debt crisis of 1982 and the subsequent stagnation of the 1980s, Mexico moved to open its economy to the world, a development that culminated in the 1994 signing of the North American Free Trade Agreement (NAFTA), signalling the decision to base its economic strategy increasingly on the growth of exports, especially to the USA (Moreno-Brid and Ros, 2009; Haber et al., 2008: Chapter 3). While these economic integration policies triggered a large increase in trade and financial inflows (Kose, Meredith and Towe, 2004: 28f),3 they did not bring the growth that might have ensured a general modernization of economic activities, strong creation of good jobs and a marked improvement in the living conditions of the population. Further, with its economy so closely tied to that of its northern neighbour, Mexico has been affected more severely than other countries of the region by the recent economic crisis (Zepeda et al., 2009). Among the factors helping to explain why, even before the recent downturn, this turning outward has been less successful than its proponents expected are: the entry of China and other low-wage Asian countries as major players in the world market, the concentration of Mexican exports in the US market and an overestimation of the growth impact of Foreign Direct Investment (FDI) (ibid.). The scale of Mexico’s growth shortfall after turning outward is reflected by the decrease in GDP growth from 3.4 percent per year in the 1990s to 1.6 percent per year over 2000–2010. Competition from China in the North American market dealt a serious blow to the growth of Mexican exports by reducing its market share (Table 5.3), a process analysed in depth by Gallagher et al. (2008). Additionally, Chinese exports to Mexico have increased, thereby accentuating Mexico’s trade deficit with Asia (ECLAC, 2011). Much of China’s increased share of US imports, doubling from around 10 percent in 2001 to almost 20 percent in 2009, occurred after it joined the World Trade Organization (WTO). In 2002, China overtook Mexico as the second largest source of US imports after Canada, and in 2007 its share exceeded that of all 18 Latin American countries together. Mexican exports did increase after the creation of NAFTA, but its share of the US market declined from 2002 to 2005 and has since then fluctuated around 10 percent (Jenkins, 2010: 2).

206

Social protection, growth and employment

Table 5.3: Latin America’s Loss of Market Share in the USA to China, 1996–2009 1996–2001

2001–2009

Mexico

- 0.9%

- 10.1%

Total Latin America

- 0.7%

- 8.2%

Source: author’s elaboration from USITC data.

The desire to attract FDI was galvanized by the aim of raising investment (Berg et al., 2006). Nonetheless, the increase in FDI flows to Mexico did not generate a large increase in the gross investment rate (which rose only from an (unweighted) average of 23.0 percent in the 1990s to 23.9 percent over the following decade); meanwhile the gross marginal capital–output ratio rose from an already high 6.6 to a very high 13.6; even over 2000–2007 growth only averaged 2.5 percent, and this ratio was at 9.3. On neither public investment nor total capital expenditure could a statistically significant effect from FDI be confirmed (Zepeda et al., 2009: 8), and the very high marginal capital–output ratio remains a puzzle and a challenge. Mexico’s new openness has made the country more vulnerable to certain external ‘shocks’ (Blecker, 2009). Analyses show that growth in Mexico has been substantially determined by three key ‘external’ variables: net financial flows.4 the real international price of petroleum and the rate of growth of the US economy. When the US crisis began to unfold in 2007, the Mexican authorities played down its dangers for the economy; only at the beginning of 2009 were more serious policies announced. Rather than major new directions, however, most were straightforward additions to the federal budget, such as investments in infrastructure, support to businesses under stress so that they would not have to dismiss personnel, and other steps designed to increase the scope of the existing Temporary Employment Programme (Programa de Empleo Temporal — PET) (ECLAC, 2010). In principle, these measures should have helped to revive economic activity, keep people in jobs and create income for the most vulnerable groups. Instead, GDP fell by 6 percent in 2009, open unemployment reached 6.4 percent in August 2009, average wages fell by 4 percent, and most new jobs were created within small companies rather than by large ones. While these policies may have had the technical potential to be effective, and despite their approval by Congress, their impacts were severely curtailed by mid-2009 reductions in federal expenditures, a result of the government’s having underestimated the effect of the crisis on its own revenues. Subsequent recovery has been gradual, and while unemployment has fallen, it remains above the pre-crisis level, wages have increased only marginally, and small businesses continue to be the main locus of job creation.

Recent trends and developments relating to labour The last few decades have seen a slow but sustained reduction in the rate of male participation in the labour market and a more marked increase in that of women (Table 5.4).

Mexico

207

Table 5.4: Labour Force Participation Rates, by Gender, 1995–2010, 2nd quarter 1995

2000

2005

2006

2007

2008

2009

2010

Men (%)

Sex

80.9

79.9

77.7

78.7

78.2

78.3

76.8

77.6

Women (%)

36.8

38.0

39.6

40.7

41.4

42.0

41.2

42.5

Source: National Employment Survey (Encuesta Nacional de Empleo (ENE) by Instituto Nacional de Estadística y Geográfica (INEGI) (second quarter of each year)..

Behind these contrasting trends is a modest decline in rates for the youngest and oldest groups of males. For women, only the youngest group has experienced a reduction, while for women aged 20–60 a significant increase occurred (Figure 5.1). Rising educational attainment has exerted downward pressure on the rates of both young men and young women. The share of males who are pensioners has risen, as has the number of inactive adult men who are neither pensioners nor disabled, a reflection of the lack of retirement and pension programmes for nonwaged workers and the relatively low coverage for waged workers. Despite the increases in the rate of female activity, its level remains far below that of such other Latin American countries as Argentina, Brazil and Uruguay, suggesting that further increases will occur in the future. A much-noted feature of Mexico’s labour market is the relatively low level of open unemployment.5 During normal periods it has usually fluctuated between 2 and 4 percent, rising during recession or crisis (Figure 5.2). Disregarding fluctuations, the rate appears to be on an upward trend since 2001. In absolute terms, the number of unemployed people rose from 935,000 in 2001 to 2.485 million in 2010.

Figure 5.1: Female Labour Force Participation Rates, by Age, 1995 and 2010 70 60 50 40 30

1995

2010

10 75+

70 to 74

65 to 69

60 to 64

55 to 59

50 to 54

45 to 49

40 to 44

35 to 39

30 to 34

25 to 29

20 to 24

14 to 19

Source: 1995: National Employment Survey, INEGI. 2010; National Survey of Occupation and Employment, 2nd semester of 2010.

208

Social protection, growth and employment

Figure 5.2: Open Unemployment Rate, 2000–2011 7 6

Percent

5 4 3 2 1 2000/04 2000/10 2001/04 2001/10 2002/04 2002/10 2003/04 2003/10 2004/04 2004/10 2005/04 2005/10 2006/04 2006/10 2007/04 2007/10 2008/04 2008/10 2009/04 2009/10 2010/04 2010/10 2011/04

Source: INEGI, Economic Information Data Bank.

These comparatively low rates might seem to be capturing little beyond frictional unemployment (the typical period it takes a person to move between jobs when these are not too hard to find). Their low level relative to the rates of some industrial countries reflects in part the fact that in Mexico open unemployment is only one of the possible ways in which the scarcity of job opportunities manifests itself; in fact, the luxury of remaining unemployed while engaging in a lengthy search for employment is limited to a small group (Myrdal, 1968: 1122–1125). Mexico’s low unemployment compared to otherwise similar developing countries remains perplexing, however, since such countries share this second feature, but most have higher open unemployment rates. The short duration of the average unemployment period (less than a month) plus the fact that in more than a third of cases the individuals have resigned or left their jobs voluntarily suggest that a significant part of the unemployment is at least partly frictional. The other cases originate mainly in discontinuations of employment, i.e. in temporary contracts that reach their end. In 2010, 65.6 percent of unemployed people were spouses or children of the head of household; in 2010, 78 percent of the individuals who left their jobs voluntarily were in these two categories. Thus a significant share of unemployed people have some family support as they search for work. As discussed further below, an important proportion of the unemployed people who get a job enter the microenterprise or ‘micro-unit’ sector, made up of units with five or fewer workers (Salas, 2003; INEGI, 2006; Junco, 2007). Employment growth has been moderate, averaging about 2 percent per year over 1992–2008. Years of very low growth in GDP (e.g. 2001, 2002 and 2009) have generally also featured low growth in employment, with renewed GDP growth showing up in high employment creation the following year. These cyclical patterns of total job creation are matched by changes in the breakdown of

Mexico

209

employment by size of establishment (Table 5.5). Over 2006–2008, that structure did not change a great deal, but as the economy confronted the crisis of 2008–2010 there was first a decline in employment in all size categories (2008–2009), though a sharper one for the large establishments; then in 2009–2010 all categories showed some recovery except the largest, but over 80 percent of the total increase occurred in microenterprises.6 Over the longer period 2005–2010, such micro-units accounted for 69 percent of all net new jobs; over 2008–2010 (including the downturn of 2009) they accounted for more than the total modest increase (since employment shrunk in the other size categories taken together (Table 5.5). The fact that open unemployment did not reach the high levels of developed countries during the crisis can be explained in considerable part by the contribution of smaller-sized units (which include self-employed work) to the job creation process. In sectoral terms, Mexico stands out for the rapid decline in agricultural employment, from over a quarter of the total in the early 1990s to 13 percent by 2010. The greatest change has been in small-scale agriculture, a sector including a significant number of home-consumption subsistence activities; it was dramatically affected by NAFTA, whose liberalization of trade (together with an overvalued exchange rate and US corn subsidies) brought massive imports of corn and beans into the country (Fox and Haight, 2010). One aim of NAFTA was to foster a sustained growth of manufacturing employment, not just in activities relating to the maquila assembly plants (Zepeda et al., 2009). Between 1995 and 2008 the increase was just over 2 million individuals, pushing the sectoral share of total employment up from 14.5 percent to 16.5 percent. Meanwhile, construction expanded strongly in percentage terms. Virtually all of

Table 5.5: Increases in Employment, by Size of Establishment, 2006–2010, 2nd quarter Year

Number of employees 1 to 5

6 to 15

16 to 100

101 or more

Total

2006

633,508

74,362

397,573

287,566

1,393,009

2006

45.48%

5.34%

28.54%

20.64%

100%

2007

304,118

101,274

23,967

156,402

585,761

2007

51.92%

17.29%

4.09%

26.70%

100%

2008

481,947

92,140

-42,594

269,020

800,513

2008

60.20%

11.51%

-5.32%

33.61%

100%

2009

-175,869

-49,676

-12,586

-501,314

-739,445

2009

23.78%

6.72%

1.70%

67.80%

100%

2010

1,070,361

45,193

314,370

-102,301

1,327,623

2010

80.62%

3.40%

23.68%

-7.71%

100%

2006–2010

2,314,065

263,293

680,730

109,373

3,367,461

2006–2010

68.72%

7.82%

20.21%

3.25%

100%

Source: Data taken from the National Survey for Employment (ENOE) by INEGI, measuring the second quarter of each year and illustrating employment increases in relation to the same quarter of the previous year.

210

Social protection, growth and employment

the net employment creation, however, came from services, with the retail trade sector, personal services and public administration responsible for the bulk of it; domestic service, traditionally an important source of employment for women, lost some ground, as female employment shifted towards trade and other services as well as manufacturing.

Characteristics of employment One of the most noteworthy trends over 1995–2010 has been the marked increase in the share of workers who are in paid employment (Table 5.6), resulting largely from a jump of almost 10 percentage points in the agricultural sector; in non-agricultural activities the increase was a modest 2 percentage points. In agriculture the absolute number of paid workers rose until the early 2000s and then fell back, leaving a small net increase between the early 1990s and 2008. As can be seen from the analysis of the National Employment Surveys, the increase in the proportion of salaried agricultural workers is in large part due to the significant drop in the number of self-employed and unpaid workers, deriving from the process of expulsion of peasants in the aftermath of NAFTA. The greatest increases in waged employment were in personal services, trade and construction. In contrast, the proportion of self-employment in manufacturing grew due to the proliferation of micro-units producing garments and foodstuffs. As noted below, these trends are, to a considerable degree, a reflection of the size of establishments, which has wide-reaching consequences for the amount of paid employment, wage and salary levels and access to social security and labour rights.

Structure of employment by size of establishment For the economy as a whole, employment is concentrated in small-scale activities, with 53 percent located in very small micro-units of up to five workers, 64 percent in units of 15 workers or less, and only 14 percent in large units of over 250 workers (Table 5.7). Micro-units play the dominant role in employment creation in agriculture, retail trade, construction and manufacturing, while large units dominate in mining and utilities (generation and distribution of energy, water and gas)—sectors that create little employment in total. Outside agriculture, the micro-units account for 43percent of employment. The proportion of waged work increases with the size of the establishment (Table 5.7). Non-waged work is thus concentrated in activities where micro-units are important, in particular agriculture, followed by retail trade and some sub-branches in the services sector. Establishment size is closely related to three key features of paid work: access to social security, access to labour rights, and job stability. It is also related to average labour income. The percentage of waged workers with access to social security7 rises from 21.6 percent for units with 1–5 workers to 95.3 percent for establishments with 251 workers or more; overall, 61.8 percent of paid workers have such access (Table 5.8). The big jump in coverage — a tripling — occurs between the establishments of up to five workers and those with 6–15.

Mexico

211

Table 5.6: Percentage of Employment by Job Position, by Economic Sector, 1995 and 2010, 2nd quarter 1995 2010 1995 2010 1995 2010 1995 2010 Employers % Agriculture, Livestock farming, Development of Forestry, Fisheries and Gamekeeping

4.1

SelfEmployed %

Waged Workers %

Non-paid Workers %

4.9

38.5

37.8

28.5

37.7

28.8

3.5

73.1

93.9

14.7

1.2

0.0 100.0

99.6

0.0

0.4

19.6

Mining

0.2

1.4

12.0

Electricity, Water and Supplying Gas through Gas Lines to the Final Consumer

0.0

Construction

9.6

10.0

23.7

18.3

65.3

71.3

1.4

0.4

Manufacturing Industries

3.8

3.9

10.9

16.1

81.3

75.7

4.0

4.3

Wholesale trade

6.4

6.3

5.2

9.5

85.2

81.5

3.2

2.7

Retail trade

4.5

4.8

43.8

40.5

32.9

43.1

18.8

11.6

Transport, Communication, Post and Storage

5.2

4.3

17.8

20.4

75.5

75.0

1.3

0.3

Information in the mass media

1.3

2.7

2.0

2.3

96.1

94.4

0.6

0.7

Financial and Insurance services

0.6

1.7

0.6

1.9

98.7

96.1

0.1

0.3

Real Estate Services and the Leasing of Movable and Intangible Assets

6.3

10.7

18.1

18.4

68.6

68.4

7.0

2.4

Professional, Scientific and Technical services

9.4

12.1

25.8

29.8

63.1

56.9

1.6

1.2

26.6

0.5

33.2

0.0

37.4

95.2

2.9

4.3

Support Services for Businesses and Waste Management and Remediation Services

1.8

2.9

43.7

16.6

49.4

78.4

5.2

2.1

Educational Services

0.3

0.6

1.7

2.3

97.9

96.4

0.2

0.7

Management of Corporations and Companies

Health and Social Welfare Services

4.9

4.6

8.5

11.1

85.6

82.2

1.1

2.1

Leisure, Cultural and Sporting Services and other Recreational Services

6.5

4.7

20.1

24.9

71.7

67.3

1.8

3.2

Temporary Housing Services and the Preparation of Foods and Beverages

7.3

7.2

24.2

27.1

55.9

54.2

12.5

11.5

Other Services except for Government Activities

3.9

4.0

27.1

21.3

66.9

73.4

2.1

1.3

Activities by the Government and International and Extra-territorial Organizations

0.1

0.0

1.1

0.0

97.7

99.0

0.5

1.0

Unspecified

0.0

1.5

0.4

1.4

77.5

83.6

0.0

13.6

Total

4.4

4.7

25.4

23.0

58.4

65.6

11.7

6.7

Non-agricultural Sector

4.5

4.7

21.3

20.7

67.7

69.8

6.4

4.7

Source: authorâ&#x20AC;&#x2122;s calculations based on ENE data for 1995 and ENOE data for 2010, as measured in the second quarter of each year by INEGI.

212

Social protection, growth and employment

Table 5.7: Employment by Economic Sector, by Size of Establishment and by Whether Paid or Not, 2010, 2nd quarter Size of Establishment (number of individuals) Agricultural activities Paid Paid/Total Mining Paid Paid/Total Electricity, Water and Gas

1 to 5

6 to 15

16 to 100

101 to 250 51,218

251 and more

Total

4,887,030

528,941

330,540

49,504

5,847,233

1,330,266

422,761

312,310

27.22%

79.93%

94.48%

50,703

49,002

2,165,042

98.99%

37.03%

18,105

14,083

38,207

7785

13,493

37,433

17,073

90,669

178,137

17,073

90,490

166,274

43.00%

95.81%

97.97%

100.00%

99.80%

93.34%

2,498

2,259

16,134

8,335

146,191

175,417

2,498

2,259

16,134

8,335

143,837

173,063

Paid/Total

100.00%

98.39%

98.66%

Construction

2,438,204

331,291

484,534

56,765

75,375

3,386,169

1,493,828

285,546

460,557

54,021

74,962

2,368,914

61.27%

86.19%

95.05%

95.17%

99.45%

69.96%

2,405,705

655,579

1,308,797

599,583

1,566,990

6,536,654

Paid

850,819

582,166

1,274,728

591,365

1,559,767

4,858,845

Paid/Total

35.37%

88.80%

97.40%

98.63%

99.54%

74.33%

Paid

Paid Paid/Total Manufacturing industries

Wholesale trade

299,710

180,581

383,564

66,771

69,814

1,000,440

Paid

136,859

155,682

371,226

65,844

69,138

798,749

Paid/Total

45.66%

86.21%

96.78%

98.61%

99.03%

79.84%

5,898,734

779,349

764,710

172,445

204,764

7,820,002

1,437,308

659,525

662,263

149,580

180,477

3,089,153

Retail trade Paid

24.37%

84.63%

86.60%

86.74%

88.14%

39.50%

1,083,967

144,515

335,285

54,605

155,720

1,774,092

Paid

640,849

129,098

326,808

54,015

154,362

1,305,132

Paid/Total

59.12%

89.33%

97.47%

98.92%

99.13%

73.57%

7,557,238 1,971,941

3,241,093

601,954

2,857,678 16,229,904

3,958,389

1,840,561

3,321,682

629,463

2,926,217 12,676,312

52.38%

93.34%

102.49%

104.57%

19,704,161 4,079,598

6,572,324

1,577,531

5,167,201 37,100,815 5,248,252 27,601,484

Paid/Total Transport, Communications, Post and Storage

Services Paid Paid/Total Total Paid Paid/Total

9,858,601

4,091,091

6,783,141

1,620,399

50.03%

100.28%

103.21%

102.72%

102.40%

101.57%

78.10%

74.40%

Source: authorâ&#x20AC;&#x2122;s calculations based on ENOE data from the second quarter of 2010. Please note that public employees are included in this table.

Mexico

213

Table 5.8: Percentage of Paid Workers with Social Security Benefits, by Economic Sector, 2010, 2nd quarter Economic Sector

Size of Establishment (Number of workers) % 1 to 5

Agriculture Mining Electricity, Water and Gas Construction Manufacturing Industries Wholesale Trade Retail Trade Transport & Communications Services All Sectors Total number of individuals Percentage of all workers having social security benefits

6.8

6 to 15 16.2

16 to 100 37.4

101 to 251 and 250 more 56.4 77.7

Total 15.7

23.3 48.3 87.3 98.0 98.2 88.3 97.3 97.0 96.6 100.0 97.6 97.7 5.9 35.2 79.0 90.4 87.8 28.0 22.3 51.1 86.9 97.6 99.0 76.8 34.9 76.8 92.8 95.3 95.7 80.9 33.6 76.4 88.9 91.5 87.8 62.3 11.0 59.9 83.0 90.8 96.6 48.1 29.1 72.3 87.8 93.9 93.9 69.1 21.6 61.6 85.1 93.8 95.3 61.8 2,119,098 2,486,898 5,664,223 1,506,490 4,927,578 16,704,287 12.7 14.9 33.9 9.0 29.5 100.0

Source: author’s calculations based on the ENOE, 2nd quarter of 2010.

Table 5.9: Percentage of Paid Workers with Labour Benefits, by Economic Sector, 2010, 2nd quartera Economic Sector

Size of Establishment (Number of workers) % 1 to 5

Agricultural activities Mining Electricity, Water and Gas Construction Manufacturing industries Wholesale trade Retail trade Transport & Communications Services All sectors Number of individuals Percentage of all workers having labour benefits

6.7

6 to 15 15.9

16 to 100 34.6

101 to 251 and 250 more 50.8 64.1

Total 15.3

23.3 47.4 83.3 94.6 96.3 85.9 94.9 92.7 82.1 96.6 96.0 94.8 5.8 32.2 70.4 81.3 75.1 26.6 21.9 49.6 83.5 95.1 96.6 74.6 34.0 74.5 90.6 95.0 94.0 78.9 32.8 74.0 87.2 90.8 87.2 60.8 10.9 57.9 78.5 87.3 94.1 46.6 29.1 70.8 85.7 92.2 92.7 68.1 21.4 59.8 82.0 91.5 93.3 60.3 2,114,972 2,485,850 5,661,140 1,506,490 4,927,477 17,417,607 12.14 14.27 32.50 8.65 28.29 100

Notes: a) According to the Secretariat for Work and Social Prevision (La Secretaría del Trabajo y Previsión Social–STPS), labour benefits are defined as “goods and services, other than access to health institutions, that workers receive in cash or in kind, on behalf of the economic unit they work for, in addition to their salary, to supplement their earnings” (see STPS glossary on labour terminology, available at: tinyurl.com/bmawqqx). Source: author’s calculations based on the ENOE, 2nd quarter of 2010.

214

Social protection, growth and employment

A similar pattern holds with respect to access to labour rights enshrined in current labour legislation, such as paid holidays, days of leave, bonuses etc. (Table 5.9). It is noteworthy that companies generally observe regulations relating to social security more than those relating to labour benefits.

Table 5.10: Percentage of Paid Workers with Labour Benefits, by Type of Labour Contract, 2010, 2nd quarter

Services

Transport and Communications

Retail trade

Wholesale trade

Manufacturing industries

Construction

Electricity, Water and Gas

Mining

Economic Sector

Agricultural activities

Size of Establishment (Number of Workers) and Type of Labour Contract 1 to 5 individuals Temporary

99.9% 100.0% 100.0%

99.8%

99.6%

99.1%

99.3%

99.8%

99.2%

Permanent

99.9%

98.8% 100.0% 100.0%

99.7%

97.4%

99.1%

99.7%

99.8%

0.0%

0.3%

1.3%

3.7%

1.7%

0.7%

1.4%

89.5% 100.0%

98.8%

98.4%

96.9%

97.1%

98.1%

92.1%

100.0% 100.0% 100.0%

99.4%

99.0%

96.2%

96.7%

98.4%

97.2%

No written contract

0.1%

1.2%

6 to 15 individuals Temporary Permanent

99.5%

No written contract

0.5%

10.5%

0.0%

1.9%

3.1%

6.9%

6.2%

3.7%

11.4%

Temporary

98.8%

94.3%

69.7%

96.0%

89.7%

88.2%

92.6%

92.2%

81.2%

Permanent

99.4%

92.9% 100.0%

98.6%

97.3%

89.6%

91.6%

97.0%

94.0%

52.2%

7.1%

14.8%

22.8%

16.0%

12.0%

26.4%

0.0% 100.0%

89.6%

75.8%

84.4%

87.6% 100.0%

67.1%

100.0% 100.0% 100.0%

96.2%

88.3%

73.6%

90.1%

92.6%

87.6%

6.6% 100.0% 100.0%

14.2%

40.7%

42.0%

23.5%

7.4%

47.9%

96.1% 100.0%

96.6%

76.8%

75.0%

92.8%

93.7%

60.4%

100.0% 100.0% 100.0%

90.2%

81.2%

84.5%

89.5% 100.0%

88.0%

16 to 100 individuals

No written contract

1.8%

12.8%

101 to 250 individuals Temporary Permanent No written contract

93.4%

251 and more Temporary Permanent No written contract

94.4% 5.6%

3.9%

0.0%

13.2%

42.2%

40.4%

19.4%

6.3%

53.1%

Temporary

97.2%

76.0%

93.9%

96.2%

88.1%

88.7%

93.9%

96.8%

80.0%

Permanent

99.9%

98.3% 100.0%

96.9%

93.1%

88.3%

93.4%

97.5%

93.3%

7.3%

20.4%

23.2%

13.4%

6.0%

28.0%

Total

No written contract

2.9%

25.7%

30.4%

Source: authorâ&#x20AC;&#x2122;s calculations based on the ENOE, 2nd quarter of 2010.

Mexico

215

The type of contract (‘temporary’, ‘permanent’ or ‘none’) also varies by sector and size of establishment, with verbal contracts prevailing in smaller units, while contracts of a more permanent nature tend to be found in larger units (Table 5.10). Having a contract is strongly related to compliance with labour regulations, both for the economy as a whole and within sectors, whereas, interestingly, the difference between having a permanent or a temporary written contract is small (Table 5.11); overall the shares of those with temporary, permanent or no written contracts having labour benefits are 90.1 percent, 95.6 percent and 17.5 percent, respectively. Although there are also significant disparities across sectors, there is a general lack of protection for workers with no written contract. For some sectors and size categories the prevalence of labour benefits is considerably less for those with temporary jobs, but for no size category is it less than 77 percent. Sometimes temporary jobs are carried out within the category of rendered professional services — a mechanism that does not require registration with labour or social security agencies. This helps also to explain the presence of self-employed workers within the larger units. In summary, the size of the production unit is a good predictor of workers’ access to the use and enjoyment of their labour and social security rights, as well as of how long they can

Table 5.11: Contractual Arrangement of Paid Workers, by Economic Sector and Size of Establishment, 2010, 2nd quarter Economic Sector/ Type of Contract

Size of Establishment (number of individuals) 1 to 5

6 to 15

16 to 100

101 to 250

251 and more

Total

Agriculture Temporary

0.2

1.2

7.9

9.6

13.1

2.02

Permanent

0.9

4.9

15.7

24.4

25.4

4.92

Without Written Contract

98.9

93.9

76.4

66.0

61.5

93.06

Total number of waged workers 1,330,266

422,761

312,310

50,703

49,002

2,165,042

0.0

8.1

21.7

34.0

30.4

25.58

Mining Temporary Permanent

5.8

16.7

54.1

60.8

65.5

55.70

94.2

75.1

24.2

5.1

4.1

18.71

7,785

13,493

37,433

17,073

90,490

166,274

Temporary

1.4

0.0

18.7

8.8

12.0

12.16

Permanent

46.2

58.1

67.5

88.9

84.6

82.31

Without Written Contract

52.3

41.9

13.8

2.3

3.4

5.52

2,498

2,259

16,134

8,335

143,837

173,063

Without Written Contract Total number of waged workers Electricity, Water and Gas

Total number of waged workers

Source: author’s calculations based on the ENOE, 2nd quarter of 2010

216

Social protection, growth and employment

Economic Sector/ Type of Contract

Size of Establishment (number of individuals) 1 to 5

6 to 15

16 to 100

101 to 250

251 and more

Total

Construction Temporary Permanent

0.4

7.6

22.3

34.8

24.8

7.08

0.8

11.7

43.9

50.9

58.0

13.45

98.8

80.6

33.8

14.4

17.2

79.46

Total number of waged workers 1,493,828

285,546

460,557

54,021

74,962

2,368,914

Without Written Contract Manufacturing industries Temporary

1.0

4.7

12.3

16.7

13.8

10.43

Permanent

5.7

23.4

60.2

76.0

83.4

55.62

93.2

71.9

27.5

7.3

2.8

33.94

591,365 1,559,767

4,858,845

Without Written Contract Total number of waged workers

850,819

582,166 1,274,728

Wholesale trade Temporary

2.1

8.3

10.5

10.1

9.7

8.53

Permanent

18.6

49.2

75.3

81.9

84.2

61.81

Without Written Contract Total number of waged workers

79.3

42.5

14.2

8.0

6.1

29.66

136,859

155,682

371,226

65,844

69,138

798,749

Retail trade Temporary

2.1

7.9

13.4

15.1

12.7

7.16

Permanent

13.1

52.9

80.5

90.6

87.1

44.54

Without Written Contract

86.1

45.5

18.7

9.1

12.4

54.29

Total number of waged workers 1,437,308

659,525

662,263

149,580

180,477

3,089,153

Transport, Communications, Post and Storage Temporary

0.9

6.9

8.1

9.2

11.5

4.89

Permanent

2.5

37.9

65.8

81.1

84.2

34.77

Without Written Contract Total number of waged workers

96.7

55.2

26.1

9.6

4.3

60.38

640,849

129,098

326,808

54,015

154,362

1,305,132

1.9

11.3

16.2

18.7

15.2

10.92

8.1

50.2

69.8

75.5

79.6

50.23

90.0

38.5

14.0

5.8

5.2

38.85

Services Temporary Permanent Without Written Contract

Total number of waged workers 3,958,389 1,840,561 3,321,682

629,463 2,926,217 12,676,312

Whole Economy Temporary Permanent Without Written Contract Total number of waged workers All sectors

9,858,601 4,091,091 6,783,141 1,620,399 5,248,252 27,601,484

Source: authorâ&#x20AC;&#x2122;s calculations based on the ENOE, 2nd quarter of 2010

Mexico

217

expect to stay in their jobs. Any lack of access to these rights for workers in larger units can be mitigated by the better pay they earn. But workers in smaller units also suffer, on average, from lower earnings (Table 5.12). In terms of the policy on employment and access to labour rights, there is a clear need to better understand the smaller units.

Table 5.12: Average Monthly Labour Income by Economic Sector, Size of Establishment and Job Position, 2010, 2nd quarter* Economic Sector/ Type of Contract

Waged workers

Employers

Self-employed workers

Agriculture 1 individual

981

2 to 5 individuals

2,203

5,118

1,069

6 to 10 individuals

2,692

5,775

1,308

11 to 15 individuals

3,465

20,849

4,047

16 to 50 individuals

3,207

11,909

930

51 and more individuals

3,722

7,663

15,289

51,314

Total

8,335

Construction 1 individual

4,017

2 to 5 individuals

3,462

6 to 10 individuals

4,452

7,387

11 to 15 individuals

5,168

10,440

16 to 50 individuals

5,699

6,702

51 and more individuals

5,489

5,152

6,383

16,738

25,164

46,756

2 to 5 individuals

2,748

5,285

2,790

6 to 10 individuals

3,216

8,180

3,765

11 to 15 individuals

3,347

5,069

1,500

16 to 50 individuals

3,466

5,376

51 and more individuals

4,516

4,439

17,293

28,349

Total

9,169

Manufacturing industry 1 individual

Total

1,944

Source: authorâ&#x20AC;&#x2122;s estimations on the basis of ENOE data, 2nd quarter of 2010.

10,049

218

Social protection, growth and employment Economic Sector/ Type of Contract

Waged workers

Employers

Self-employed workers

Trade 1 individual

2,386

2 to 5 individuals

2,721

5,623

3,538

6 to 10 individuals

3,697

6,431

4,786

11 to 15 individuals

4,176

14,153

361

16 to 50 individuals

4,262

11,240

51 and more individuals

4,652

39,132

19,508

76,579

Total

11,071

Services 1 individual

2,145

2 to 5 individuals

3,111

3,358 6,341

3,673

6 to 10 individuals

4,119

9,616

7,063

11 to 15 individuals

4,643

10,765

3,220

16 to 50 individuals

4,822

9,670

4,606

5,708

20,373

24,548

56,765

21,920

51 and more individuals Total Whole Economy 1 individual

2,145

12,686

2 to 5 individuals

14,245

27,856

16,222

6 to 10 individuals

18,176

37,389

16,922

11 to 15 individuals

20,799

61,276

9,128

16 to 50 individuals

21,456

44,897

5,586

51 and more individuals

24,981

88,345

101,802

259,763

60,544

Total

* Numbers in the third column apparently showing the presence of self-employed workers in larger units refer to individuals who work for fees on specific projects but are who are considered to be self-employed to avoid additional labour costs.

General characteristics of micro-units In terms of employment policies and access to labour rights, smaller units play a crucial role in ensuring the survival of a significant group of workers or securing an entry point for small-scale capitalist activities. It is thus worthwhile examining the functioning of small-sized units and their changing role in the economy and the labour market. Moreover, employment policies and policies to grant access to labour rights cannot be applied without a careful study of this type of unit. The ENAMIN 2008 survey permits an in-depth look at these units, their owners and their workers. This source defines microenterprises as units of up to five workers in the trade and

Mexico

219

services sectors and up to 15 workers in manufacturing. It further distinguishes between three categories: those employing waged workers, self-employed workers accompanied by unpaid family workers, and self-employed workers operating on their own.

Table 5.13: Number of Microbusinesses by Whether the Owner Hires Paid Workers and by Number of Workers, 2008 Position of Owner Employer

Self-employed*

Total

Total Businesses

1,044,460

7,064,295

8,108,755

Businesses with workers

1,044,460

1,488,640

2,533,100

1 worker

474,355

1,059,086

1,533,441

2 workers

281,376

276,295

557,671

3 workers

146,825

115,593

262,418

4 workers

78,225

28,130

106,355

5 workers

39,715

2,889

42,604

6–15 workers

10,845

1,024

11,869

Unspecified

13,119

5,623

18,742

* Workers in units operated by ‘self-employed’ owners are unpaid family workers. Source: National Microbusinesses Survey 2008.

Microenterprises are notoriously heterogeneous; those more orientated to or successful in achieving growth and accumulation tend to employ waged labour, while others are primarily oriented towards subsistence and the social reproduction of their families (Picchio, 1992); some units that begin closer to the latter modality transition to the former. Whether unpaid family helpers are involved depends a good deal on the setting. The range of types of micro-units creates a need for public support policies that are customized by segment. ENAMIN reports the presence of over 8 million microenterprises in 2008, with just over 1 million having paid workers. Total employment was about 11 million workers, about 8 million with no paid workers, and the rest with paid workers.8 According to the ENAMIN survey, the main direct reason for beginning a small-scale activity is a quest for income (46 percent, if we include the categories ‘complement family income’ and ‘more money than as a wage earlier’), while less than a quarter (22.7 percent) seemed in their responses to be alluding to ‘entrepreneurship’ (found a good ‘business opportunity’ and ‘wanted to be independent’), and 8 percent referred to carrying on the family tradition (Table 5.14). In any case, it is clear that the microenterprise sector is not made up primarily of ‘entrepreneurs’ who start a business even though they have other good alternatives but also includes individuals seeking income for reasons of survival and related to the general state of the economy. Data shown in Table 5.14 also imply that, though important to some (less than 10 percent of the

220

Social protection, growth and employment

Table 5.14: Main Reasons for Beginning a Small Business, 2008 Number Family tradition

Percentage

648,948

8.00

To complement family income

2,612,963

32.22

For more money than as wage earner

1,119,426

13.81

Did not find employment as a waged worker

562,650

6.94

Flexible timetable

218,582

2.70

Dismissal or laying off of personnel

139,068

1.72

Wanted to be independent Found a good business opportunity

603,600

7.44

1,239,919

15.29

Other

963,599

11.88

Total

8,108,755

100.00

Source: National Microbusinesses Survey 2008.

Table 5.15: Microbusinesses, by Economic Activity, by Presence of Paid Workers or not, and by Age, 2008 5 years and less (%) Total Manufacturing

6 years and more (%)

42.7

57.3

33.1

66.9

Construction

25.4

74.6

Trade

49.8

50.2

Services

43.0

57.0

Transport

38.6

61.4

33.1

66.9

Employer Manufacturing

33.0

67.0

Construction

22.4

77.6

Trade

32.1

67.9

Services

36.9

63.1

Transport

40.4

59.6

Self-employed worker

44.1

55.9

Manufacturing

33.1

66.9

Construction

26.5

73.5

Trade

51.3

48.7

Services

44.1

55.9

Transport

38.4

61.6

Source: National Microbusinesses Survey 2008.

Mexico

221

sample), the desire for independence or flexibility is not the dominant motivator for engaging in these kinds of activities. The success of microenterprise units, measured in terms of longevity, is uneven, with 57 percent of those surveyed having been in existence for a substantial period of time (six years or more) (Table 5.15). Average completed life of these firms would, by definition, be greater than their average life at the time of the survey, suggesting a perhaps considerable longevity.9 This finding, consistent with those of all ENAMIN surveys since 1992, leaves no doubt that microenterprises are not primarily temporary activities in which workers take refuge, pending their return to jobs in larger-scale units.10 As will be noted in Section 5, the flow of workers between micro- and medium- or large-scale units is negligible. Despite their survival capacity, micro-units face many problems. Most frequently cited is ‘low sales or excessive competition’, which reflects the first or direct level at which firms perceive their difficulties. Other factors, such as high or rising costs, sometimes lie behind the

Table 5.16: Main Problems Reported by Microenterprises, by Availability of Fixed Location for Business, 2008 (percentage citing the problem)* No fixed location (%)

Fixed location (%)

Total (%)

Excess of debt, taxes or high interest

0.49

1.82

0.95

Rising costs

9.28

5.04

11.09

46.21

48.18

46.90

Lack of credit or capital

Low sales or excessive competition

2.35

2.62

2.44

Noncompliance of suppliers or poor quality of inputs

0.20

0.33

0.24

Conflicts with workers

0.00

0.09

0.03

Lack of training

0.11

0.17

0.13

Excessive bureaucracy

0.15

0.42

0.24

Problems relating to payment by customers

0.41

1.97

2.16

Problems with authorities or public insecurity

2.10

2.26

2.16

17.25

13.13

15.82

Others No problems evident Total number of firms

21.46

14.54

19.05

5,162,976

2,755,717

7,918,693

* The distinction between the categories ‘fixed location’ and ‘no fixed location’ is based on the availability of specific sites and facilities for the execution of business activities. Whereas the latter encompasses vehicles with or without motor, improvised or fixed spots on public roads, markets or outdoors, semi-fixed selling activities from door to door and in the homes of clients, the former refers to more formalized locations, such as shops, which are installed on a market under a common roof, in a shopping mall, in a factory or an office (ENAMIN, 2008: 63 and 239). Source: National Microbusinesses Survey 2008.

222

Social protection, growth and employment

market problem. It is striking that relatively few units refer to problems related to lack of credit or working capital or to problems with bureaucracy. As expected, the interface between these very small firms and the agencies of the state is limited. Registration with local, state or federal authorities is a normal legal requirement to conduct an economic activity, as is registration with both the Municipality and the Health Department. Access to support programmes requires registration with the Economy Secretariat. Registration is uncommon, however; the most frequent type (with the Municipality) is undertaken by less than a quarter of these establishments (Table 5.17), while membership in a trade association is even less common, at about 10 percent (Table 5.18). As for health services, only a modest 20 percent of the micro-unitsâ&#x20AC;&#x2122; waged workers are enrolled in the IMSS, along with under 10 percent of the self-employed workers (Table 5.19). For the self-employed, the Community (Popular) Insurance, a support mechanism for non-waged

Table 5.17: Type of Registration with an Institution by Microenterprises, 2008 Total

Type of registration Municipality or delegation

Economy Secretariat

Health Department

None of these

Total

8,108,755

1,963,281

250,137

338,059

5,863,202

Employer

1,044,460

453,798

99,705

118,597

493,865

Self-employed worker

7,064,295

1,509,331

150,432

219,462

5,369,337

Source: National Microbusinesses Survey 2008.

Table 5.18: Trade Union Membership by Type and by Type of Microenterprise, 2008

260,719

173,868

143,678

143,167

129,516

7,257,807

Employer

1,044,460

38,151

23,872

48,735

42,335

33,584

857,783

Selfemployed worker

7,064,295

222,568

149,996

94,943

100,832

95,932

6,400,024

Source: National Microbusinesses Survey 2008.

None

8,108,755

Other

Chamber of Commerce or Industrial

Total

Professional association

Transport workers, Taxi drivers

Type of registration Tenants, stallholders etc.

Total

Mexico

223

Table 5.19: Access to Health Services, by Type of Microenterprise and Type of Health Service, 2008 Position at work

Total

Type of registration IMSS

Private medical service

‘Community’ insurance

Other

10,129,994

12,283,418

1,348,587

143,637

625,342

Employers

1,044,460

214,262

22,343

53,725

754,130

Self-employed workers

7,064,295

672,145

83,676

550,736

5,757,738

480,473

21,538

2819

1023

5250

449,843

Subordinate waged workers*

1,938,979

431,133

28,826

12,471

24,484

1,442,065

Non-waged workers

1,755,211

9509

5973

7387

6124

1,726,218

Partners

35,858

None

* Subordinate waged workers are persons who have served as wage/paid and non-wage/non-paid workers. - Waged or paid worker: a person who performs any productive activity and in return receives remuneration via a salary or wage. - Non-waged or non-paid worker: a person who performs any productive activity to complement their or their family’s livelihood without receiving in return any remuneration via monetary flows or goods, subject to market exchange. Source: National Microbusinesses Survey 2008.

workers outside the IMSS, brings another 8 percent under some form of coverage. The ENAMIN figures demonstrate the general lack of access to medical services—including basic services— among individuals employed in microbusinesses. Among the groups distinguished in Table 5.19, coverage of one sort or another ranges from 1.7 percent for unpaid family helpers to a maximum of 28 percent for the employers.

The dynamics of labour activity There is much movement of individuals between inactivity, unemployment and activity, as evidence by figures covering the period between the 2nd and 3rd quarter of 2010 (recall that 2010 was a year of employment recovery after the overall decline in 2009). Over this three-month period, a full 15 percent of those originally employed changed status, with 5.2 percent becoming unemployed or inactive but available (persons who are not actively pursuing an economic activity but are willing to work if an opportunity arises), and nearly 10 percent moving to more definite inactivity (Table 5.20). Nearly a quarter of those initially unemployed had made the transition to employment, while a little under a quarter stayed unemployed, and over a quarter moved to inactivity, presumably in part because of discouragement. Many fewer of those initially inactive became employed (about 15 percent), with most staying in the same category.

224

Social protection, growth and employment

Table 5.20: Flows between Employment, Unemployment and Economic Activity, 2010, 2nd and 3rd quarters (persons 14 years and older) State of Activity in Second Quarter, 2010

State of Activity in Third Quarter, 2010 (%) Employed

Unemployed

Employed

85.4

2.7

Unemployed

48.5

Inactive available

18.2

Inactive not available

13.4

Inactive available

Inactive not available

2.5

9.5

23.6

8.4

19.6

4.3

20.5

57.0

1.8

10.8

74.0

Source: author’s calculations basis on ENOE data (second and third quarters of 2010).

The great majority of those who made the transition from unemployment or inactivity to employment took up jobs in trade or personal services, and for nearly three quarters (73.7 percent) the employment was in a business of five or fewer workers. Since at this time only a little over half of all workers were in such units (Table 5.7), the receiving capacity of these small firms was thus about three times greater than that of larger firms.11 Those who remained employed between the second and third quarters of 2010 also experienced considerable movement across establishment sizes (Table 5.21). Thus of those initially working alone, one third were, three months later, working in a larger unit, with 27.6 percent moving to the 2–5 worker category, and 7.0 percent to larger firms. Symmetrically, a third of those starting in establishments of over 250 workers wound up in a smaller one, with most of these (22.9 percent) still in fairly large units (16–250 workers), and a few (7.1 percent) in a smaller one between fairly adjacent size categories. The observed movements of people between size categories are, however, the combined result of

Table 5.21: Distribution of Employed Persons by Size of Establishment, 2010, 2nd and 3rd quarter Number of individuals

2–5

6–10

11–15

16–50

51–100 101–250

251 and more

65.40

27.60

2.30

0.60

1.70

0.70

0.30

0.90

2 to 5

15.30

69.20

7.40

1.70

2.90

0.90

0.50

1.10

6 to 10

5.80

34.00

26.80

8.60

14.30

3.30

1.10

2.90

11 to 15

4.00

16.50

20.20

18.00

28.20

4.30

1.70

3.20

16 to 50

3.10

9.50

8.80

9.20

39.20

11.60

5.80

8.40

51 to 100

1.50

6.40

4.40

3.20

25.30

21.20

14.40

19.10

101 to 250

1.70

3.70

2.30

2.00

15.90

15.00

24.00

31.30

251 and more

1.60

2.80

1.60

1.10

7.30

6.50

9.10

67.30

Source: author’s calculations basis on ENOE data (second and third quarters of 2010).

Mexico

225

some people changing firms together with the movement of the enterprises themselves between size categories, and the source does not permit a distinction between these two types of mobility. It seems likely that the pattern of movement between adjacent sizes holds both for people who change firms and for firms that change size categories. In either case, movements between segments distant from each other—for example, from micro-units towards medium-sized or large enterprises—are a relatively rare phenomenon.

Policies for the very small economic units A variety of specific policies have been created to improve the competitiveness of micro and small firms by improving their access to resources and providing training programmes for employers and workers. The Small and Medium-Sized Enterprises Fund of the Economy Secretariat (Fondo PyME), founded in 2004, is designed to provide “support to projects that encourage the creation, development, consolidation, viability, productivity, competitiveness and sustainability of micro, small and medium-sized enterprises…”. Its resources, originating with the federal and local governments and the private sector, are managed by intermediary entities established to select which projects to fund. The intermediaries can be financial or educational institutions, business organizations, development banks, state governments or civil associations, as long as their aims are compatible with at least one of the Fund’s objectives. The Fund is directed at five business segments: entrepreneurs, microenterprises, small and medium-sized enterprises (SMEs), ‘gazelle’ companies (those with a potential for sustained high growth) and driver or magnet companies (whose presence can attract other firms to start up in the same vicinity). These segments receive support for finance, trading, training and consulting services (Fund Evaluation, 2007–2008).

Table 5.22: Support Programmes for Micro-units, Known of and Used, 2008 Programme National Programme for Microenterprise Financing (PRONAFIM) Training Support Programme (PAC)

Aware Of

Used

806,815

25,023

347,785

10,536

1,074,250

49,959

60,284

133

Microfinancing Fund for Rural Women (FOMMUR)

238,442

11,244

Programme for Modernizing Tortilla Industry (MI TORTILLA)

188,407

2103

GeneraciónUno y Generación Dos (Generation One and Generation Two) (PIPS1 and PIPS2)

156,773

4896

National Entrepreneurs Programme/National Network for Sponsoring Microcompanies

169,578

5489

Support Fund for Micro, Small and Medium-sized Companies (Fondo PyME) Competitive Programme in Logistics and Storage Centres (PROLOGYCA)

Source: National Microbusinesses Survey 2008.

226

Social protection, growth and employment

ENAMIN 2008 data and other evaluations of the Fund show its limited impact in this sector (Table 5.22), with just 4.7 percent of units availing themselves of such support. This case is not the exception, with Table 5.22 pointing to a comparably small impact of other national programmes designed to support micro-units. The National Programme for Microenterprise Financing (PRONAFIM) aims to promote the productive initiatives of individuals and social groups in poverty conditions in every region of the country by encouraging and promoting a micro-financing system with the participation of microfinance institutions (MFIs) as channels of distribution and execution of the credit, with special emphasis on those with higher poverty levels (UNAM, 2012). Evaluations (ibid.) show that this programme helps to increase income in those units that received credit from it, although job creation was small or negligible. For both PRONAFIM and Fondo PyME, the two programmes that are used the most, coverage seems unrelated to the age of the enterprise. Data are available (from the National Survey of Occupation and Employment –ENOE) on the coverage of social programmes, such as community medical insurance, and on economic support for micro-units, such as training scholarships and other forms of monetary support, where the assistance cannot be attributed to a specific programme; they complement and provide a crosscheck on the estimates from governmental administrative records. For the several governmental

Table 5.23: Access to Community Health Insurance, by Whether Employed or Not and, for Those Employed, by Decilea of Per Capita Family Income,b 2010, 2nd quarter Programme

Yes

Not employed

7,641,369

Yes/Total(%)

34,763,157

18.0

Employed, by per capita family income decile Decile 1

884,578

4,616,515

16.1

Decile 2

860,975

3,109,085

21.7

Decile 3

759,986

3,220,404

23.1

Decile 4

666,783

2,975,890

18.3

Decile 5

773,347

3,508,660

16.6

Decile 6

672,176

3,386,471

13.8

Decile 7

448,582

2,807,949

13.2

Decile 8

621,920

4,100,236

10.6

Decile 9

404,747

3,399,140

10.6

Decile 10 Total

397,506

3,623,329

9.9

14,131,969

69,510,836

16.9

a) This table includes all individuals of age 14 and older. Unemployed and economically inactive people are included together under the category ‘Not Employed’. b) Information was not reported for a small number of families (0.6 percent of the total). Source: ENOE (second quarter of 2010)

Mexico

227

programmes aiming to provide training grants or support for microenterprises, the figures reported in the government reports do not coincide with those derived from the ENOE survey. This might be explained by the usage of sampling methods in the case of the latter, with the survey not capturing complete beneficiary populations but reconstructing totals via the design of samples. The Community Health Insurance (Seguro Popular de Salud) medical insurance scheme for non-economically active and self-employed workers who do not have access to either public or private social security (IMSS, ISSSTE) was launched as a pilot project in 2001 and institutionalized in 2004. Available data from ENOE (Table 5.23) imply that about 18 percent of all nonworking people accessed the service in 2010. When the numbers of employed people with access to this programme are tabulated according to the family decile of labour income, a progressive impact becomes visible: the degree of access by upper deciles is striking, with the share of people availing themselves falling from a peak of 23.1 percent in the second decile to 9.9 percent in the top decile. Also noteworthy is the fact that the bottom decile took less advantage of the service than did the second decile, possibly for logistical reasons (distance from a clinic etc.). Of particular interest are the 35 federal programmes aiming to promote projects that generate employment and income. The diversity of these programmes makes an overall evaluation virtually impossible. In addition, some of the programmes do not provide an adequate evaluation of coverage in terms of the population potentially capable of being served and the number of beneficiaries. Therefore, we will only examine a selection of the existing programmes. In addition to the support programmes for specific groups of producers, such as Fondo PyME, there is a group of programmes more closely geared towards creating short-term employment. Of special interest is the Temporary Employment Programme (PET), which aims to provide temporary income support for men or women during periods of low labour demand and natural or economic emergencies, with a focus on less urbanized areas where employment opportunities and infrastructure are less adequate, and which offers a wage low enough to limit participation to workers in extreme poverty. As has been shown in Coneval evaluations (Coneval,

Table 5.24: Coverage of the Employment Support Programme (PAE), 2003â&#x20AC;&#x201C;2010 Year

Estimates of potential target population

Programme objective

Coverage

2003

N/A

272,562

280,051

2004

N/A

220,765

243,195

2005

N/A

227,510

326,439

2006

N/A

207,158

244,146

2007

N/A

207,158

245,621

2008

20,917,806

261,277

403,184

2009

N/A

5,000,000

385,454

2010

N/A

6,940,131

439,842

Source: La SecretarĂa del Trabajo y PrevisiĂłn Social (STPS).

228

Social protection, growth and employment

2008), while useful, the temporary nature of the support means that neither the individuals nor the communities included in the programme are likely to emerge from the conditions of social marginalization in which they find themselves. Such short-term subsistence support programmes are—by their very nature—not designed to substantially transform the self-perpetuating social problems confronting the localities where they are implemented. Another national initiative of considerable scope is the Employment Support Programme (PAE), designed to improve the labour market by facilitating “the placement in a job or productive activity of unemployed or underemployed people, by means of the granting of support measures, either economic or in kind, for training, self-employment, labour mobility or assistance to returning emigrants” (Subsecretaría de Empleo y Productividad Laboral, 2011: 2). Its objective is to improve the interface between employers and job seekers in response to the perceived lack of: i) information regarding available job seekers and existing vacancies; ii) resources needed to look for or keep a job, to initiate or to strengthen a self-employed activity or to relocate to markets with a supply of jobs; and iii) adjustment of the labour skills of the workers. As Table 5.24 shows, its coverage is very limited: in 2009 only 385,000 individuals took advantage of it, out of a target population of 5 million people.

Conclusion The Mexican case study provides a detailed picture of the relationship between firm size, sector, type of work (permanent vs. temporary), on the one hand, and worker benefits and the reach of several government programmes, on the other. It is clear that only a minority of Mexican workers could be described as having adequate working conditions, including a modicum of security, and that a large share of new jobs over recent years have been created in microenterprise. Recent trends lead many observers to the conclusion that the informal sector is likely to remain large in Mexico for many decades; it follows that policy decisions relating to the sector and the enormous number of people who live and work in it will be central to the welfare of many Mexicans over the decades to come. This chapter highlights, among other things, the heterogeneity of the small-enterprise and informal sectors (overlapping but not synonymous with each other), not only in the nature of the productive activity in which people engage but also in the degree and type of interface with public institutions and the types of social protection available to the workers. This heterogeneity provides the setting for two of the great policy challenges facing Mexico and (either now or in the future) most other developing countries: • establishing efficient social programmes with wide coverage; and • establishing efficient support programmes to raise productivity and incomes in smaller/ more informal enterprises. It is clear that a policy focus too narrowly directed towards economic growth and the expansion of the formal sector is unlikely to resolve the problems of poverty and inadequate

Mexico

229

employment of Mexicans in the foreseeable future. A healthy growth policy can, however, be consistent with these two other objectives; achieving this consistency will be the central test of policy for some time in Mexico.

Notes 1 When an economy rebounds from recession, recovery in the labour market often follows that of output with a lag. Severe financial crises are typically associated with profound declines both in output and employment. From their study of such crises around the world, Reinhard and Rogoff (2009) conclude that, whereas the unemployment rate rises an average of 7 percentage points over the down phase of the cycle, which lasts on average over four years, output falls an average of over 9 per cent, but over a shorter period. In this light, the sluggishness of the global labour market is, in many respects, to be expected in the middle of the crisis, even though global economic growth has rebounded somewhat. 2 The ‘good intentions, bad outcomes’ thesis argues that Mexico has a dual system of social insurance that inadvertently subsidizes informal workers and firms. According to this thesis, some of these social programmes significantly contribute to hampering growth, creating low-productivity, informal jobs, fostering illegality, trapping many in poverty (see Levy, 2008). 3 Mexico’s exports to the USA and Canada tripled in US dollar terms between 1993 and 2002. While the growth of trade has slowed since 2000, Mexico’s trade (exports plus imports) with NAFTA partners were still equal to around 40 percent of its GDP in 2002. Similarly, NAFTA helped boost Foreign Direct Investment (FDI) flows to Mexico, which rose from US$12 billion over 1991–1993 to roughly US$54 billion in 2000–2002 (Kose, Meredith and Towe, 2004: 28). Increased trade and financial linkages have affected the dynamics of economic growth in Mexico in several ways. Contributions of exports and foreign investment to GDP growth have increased substantially following the introduction of the agreement. For example, the ratio of foreign investment to GDP growth reached 3 percent during 1996–2002 (ibid.: 29). 4 The net transfer of financial resources measures the total receipts of financial and other resource inflows from abroad and foreign investment income minus total resource outflows, including increases in foreign reserves and foreign investment income payments. 5 Unemployment is defined in accordance with the recommendations of the International Labour Organization (ILO) (see ILO, 1982: para. 10). 6 Note that the changing distribution of employment by establishment size reflects not only entry, exit and establishment growth in a given category but transitions of some establishments from one category to another. Thus part of the decrease in the largest category over 2008–2009 was no doubt the result of the size of some establishments falling below 100 workers. 7 That is, medical services and pensions that are the legal right of workers under contract and registered through the Instituto Mexicano del Seguro Social (IMSS), the Instituto de Seguridad y Servicios Sociales de los Trabajadores del Estado (ISSSTE) or other state institutions. 8 Figures derived from Table 5.11, with approximations for the categories 6–15 and ‘unspecified’ and the assumption that 80 percent of the employers also worked in the microenterprise. A comparison with Table 5.8, where the number of people employed in units of 1–5 people outside agriculture in 2010 was 14.817 million (to which should be added 656,000 in manufacturing with 6–15 workers) suggests,

230

Social protection, growth and employment as is to be expected, that the survey missed a significant number of enterprises, presumably mainly smaller, more hidden (e.g. those working from home) and more transient ones.

9 On the other hand, the firms missed in the survey would almost certainly have a shorter average life than those captured. 10 For an in-depth analysis, see Perry et al. (2007). 11 Disregarding the probably small number who moved out of employment and back into it over this period.

Mexico

231

References Aspe Armella, P., El camino mexicano de la transformación económica [The Mexican Path to Economic Transformation], Fondo de Cultura Económica, Mexico, 1993. Blecker, R.A., ‘NAFTA, The Peso Crisis, and the Contradictions of the Mexican Economic Growth Strategy’, Center for Economic Policy Analysis Working Paper, No. 3, Series I: Globalization, Labour Markets, and Social Policy, Center for Economic Policy Analysis, New York, July 1996. Blecker, R.A., ‘Comercio, empleo y distribución: efectos de la integración regional y global’, El Colegio de México, Los grandes problemas de México; v. 9: Crecimiento económico y equidad [‘Trade, employment and distribution: the effects of regional and global integration’, The Mexican School. The major problems facing Mexico, v. 9: Economic growth and equity], El Colegio de México, Mexico: 175–214. Blecker, R.A., ‘External shocks, structural change, and economic growth in Mexico, 1979–2007’, World Development, 2009, 37(7). Carnoy, M., ‘Race, Gender, and the Role of Education in Earnings Inequality: An Introduction’, Economics of Education Review, 1996, 15(3): 207–212. Centro de Estudios Latinoamericanos (CESLA), Economic Outlook Mexico, CESLA, Madrid, 2012. Comisión Nacional de Evaluación (Coneval), Informe de evaluación de la política social en México, 2008, Coneval, Mexico, 2008. Economic Commission for Latin America and the Caribbean (ECLAC), La República Popular China y América Latina y el Caribe: hacia una nueva fase en el vínculo económico y comercial [The People’s Republic of China and Latin America and the Caribbean: towards a new phase in economic and commercial ties], ECLAC, Santiago, 2011. England, P., Comparable Worth: Theories and Evidence, Aldyne de Gruyter, New York, 1992. Fernández, O., ‘La restricción por balanza comercial al crecimiento’, El Colegio de México, Los grandes problemas de México; v. 9: Crecimiento económico y equidad [‘Restrictions due to commercial checks and balances on growth’, The Mexican School. The major problems facing Mexico; v.9: Economic growth and equity], El Colegio de México, Mexico, 2010: 133–174. Foley, D. and T. Michl, Growth and Distribution, Harvard University Press, Cambridge, MA, 1999. Galbraith, J.K., Created unequal: the crisis in American pay, The University of Chicago Press, Chicago, IL, 2000.

232

Social protection, growth and employment

Gallagher, K., J.C. Moreno and R. Porsekansky, ‘The Dynamics of Mexican Exports: Lost in (Chinese) Translation’, World Development, 2008, 36(8): 1365–1380. Ghiara, R. and Zepeda, E., ‘Determinación del Salario en México: 1987–1993’, Economía, Sociedad y Territorio II [‘Determining salary in Mexico: 1987–1993’, Economy, Society and Territory II], 1999, 5, January–July. Gordon, D., Fat and mean: The corporate squeeze of working Americans and the myth of managerial “downsizing”, Simon and Schuster, New York, 1996. Haber, S., H.S. Klein, N. Maurer and K.J. Middlebrook, ‘Mexico Since 1980’, The World Since 1980, Cambridge University Press, Cambridge, 2008. Inter-American Development Bank (IDB), ‘IDB approves second phase of financing for Mexico’s Programa Oportunidades. Renewed support for successful program benefiting 5 million poor families’, News Release, IDB, Washington, DC, 21 March 2005, www.tinyurl.com/ c3dud8g (accessed 17 Dec. 2012). International Labour Organization (ILO), Resolutions Concerning Economically Active Population, Employment, Unemployment and Underemployment Adopted by the 13th International Conference of Labour Statisticians, ILO, Geneva, 1982. Instituto Nacional de Estadística y Geográfica (INEGI), ‘Encuesta Nacional de Empleo (ENE)’ [‘National Employment Survey’], INEGI website, Mexico, 2012, www.tinyurl.com/clu5qq7 (accessed 17 Dec. 2012). Jenkins, R., ‘Chinese Competition, Latin American Exports and the Economic Crisis’, prepared for the 2010 Congress of the Latin American Studies Association, Toronto, 6–9 October 2010, School of Development Studies, University of East Anglia, Norwich, 2010. Draft—not be quoted without permission, R.O.Jenkins@uea.ac.uk . Knaul, F.M., E. González-Pier, O. Gómez-Dantés, D. García-Junco, H. Arreola-Ornelas, M. Barraza-Lloréns, R. Sandoval, F. Caballero, M. Hernández-Avila, M. Juan, D. Kershenobich, G. Nigenda, E. Ruelas, J. Sepúlveda, R. Tapia, G. Soberón, S. Chertorivski and J. Frenk, ‘The quest for universal health coverage: achieving social protection for all in Mexico’, The Lancet, 2012, 380(9849): 1259–1279. Kose, M.A., G.M. Meredith and C.M. Towe, ‘How Has NAFTA Affected the Mexican Economy? Review and Evidence’, IMF Working Paper, International Monetary Fund (IMF), Washington, DC, 2004. Levy, S., Good Intentions, Bad Outcomes: Social Policy, Informality and Economic Growth in Mexico, Brookings Institution Press, Washington, DC, 2008.

Mexico

233

Lora, E. and J. Fajardo, ‘Latin American middle classes: the distance between perception and reality’, Interamerican Development Bank Working Paper, No. 275, Washington, DC, 2011. Moreno-Brid, J.C. and J. Ros, Development and growth in the Mexican economy: a historical perspective, Oxford University Press, Oxford, 2009. Moreno-Brid, J.C. and J. Ros, ‘El desarrollo a largo plazo de la economía mexicana: 1810–2008’ [‘Long-term development of the Mexican economy: 1810–2008’] in O. Kacef and P. Gerchunoff (eds), Institucionalidad y desarrollo económico en América Latina [Institutionality and Economic Development in Latin America], ECLAC, Santiago, 2011. Moreno-Brid, J.C., J.E. Pardinas and J. Ros, ‘Economic development and social policies in Mexico’, Economy and Society, 2009, 38(1) February: 154–176. Myrdal, G., The Asian Drama. An Inquiry into the Poverty of Nations, A 20th Century Fund Study, Pantheon, New York, 1968. Reinhard, C. and K. Rogoff, ‘The Aftermath of Financial Crises’, NBER Working Paper, No. 14656, National Bureau of Economic Research, Cambridge, MA, 2009. Salas, C., ‘Empleo y Salarios en México, 1995–2007’ [‘Employment and Salaries in Mexico, 1995–2007’] in P. Baltar, J. Dari Krein and C. Salas (eds), Economia e Trabalho: Brasil e Mèxico [Economy and Labour: Brazil and Mexico], LTr Editora, São Paulo, 2009: 170–198. Salas, C., ‘Social structures of accumulation and the condition of the working class in Mexico’ in T. McDonough et al. (eds), Contemporary capitalism and its crises, Cambridge University Press, Cambridge, 2010: 286–308. Salas, C. and E. Zepeda, ‘Employment and Wages: Enduring the Cost of Liberalization and Economic Reform’ in K.J. Middlebrook and E. Zepeda (eds), Confronting Development: Asssessing Mexico’s Economic and Social Policy Challenges, Stanford University Press, Stanford, CA, and the Center for US–Mexican Studies, University of California, San Diego, CA, 2003. Subsecretaría de Empleo y Productividad Laboral, Manual de Procedimientos del Programa de Apoyo al Empleo, Supervisión y Contraloría Social, Guía Operativa 2011 [Procedures Manual for the Employment Support, Supervision and Social Control Programme, Operational Guidance 2011], Secretaría del Trabajo y Previsión Social (STPS), Mexico, 2001. United Nations, The MDGs Report 2011, United Nations, New York, 2011. United Nations Department of Economic and Social Affairs (UN DESA), Financial flows to developing countries, UN DESA, New York, 2011.

234

Social protection, growth and employment

Universidad Nacional Autónoma de México (UNAM), Evaluación de impacto 2008–2009. Programa Nacional de Financiamiento al Microempresario (PRONAFIM). Reporte Final [Impact evaluation 2008–2009. National Programme for Microenterprise Financing (PRONAFIM). Final Report], UNAM, Mexico, 2012, www.pronafim.gob.mx/documents/doc_12_40_341.pdf (accessed 30 Nov. 2012). USAID, ‘Financing rural finance institutions in México’, MicroREPORT, No. 74, USAID, Washington, DC, 2007, pdf.usaid.gov/pdf_docs/PNADI465.pdf (accessed 17 Dec. 2012). Wolff, E., Does Education Really Help?, Oxford University Press, Oxford, 2006. World Bank, ‘World Bank Database, Country Overview Mexico’, World Bank website, 2012, data.worldbank.org/country/mexico (accessed 17 Dec. 2012). Zepeda, E., D. Alarcón, F. Veras Soares and R. Guerreiro Osorio, ‘Growth, Poverty and Employment in Brazil, Chile and Mexico’, IPC-IG Working Paper, No. 42, International Policy Centre for Inclusive Growth, Brasilia, 2007. Zepeda, E., T.A. Wise and K.P. Gallagher, Rethinking Trade Policy for Development: Lessons From Mexico Under NAFTA, Carnegie Endowment for International Peace Policy Outlook, Washington, DC, 2009.

6 Guaranteeing Jobs for the Rural Poor: An Assessment of India’s MGNREGA public works programme Eduardo Zepeda, Diana Alarcon, Scott McDonald, Chandan Sapkota, Manoj Panda, and Ganesh Kumar Introduction On 25 August 2005 the Indian parliament enacted a law guaranteeing the right of rural households to a minimum of 100 days of paid work; this important piece of legislation was later renamed the Mahatma Gandhi National Rural Employment Guarantee Act (NREGA).1 In many ways, it represents a milestone in the history of employment generation programmes in India but perhaps also a point of reference for social policy in many other developing countries. Its rights-based approach, social inclusion features, reliance on local self-government and focus on livelihoods make it an important public endeavour. Its size has no precedent in India or in any other country, posing important design and management challenges. The NREGA mandated the implementation of an ambitious, demand-driven employment creation programme aiming to benefit poor people in rural areas directly through the income provided by jobs paying a socially acceptable wage and indirectly through the execution of investment projects that improve productivity in agriculture and alleviate land degradation. The Act also set important social goals, including women’s empowerment and improved opportunities for marginalized groups, and seeks to strengthen community decision-making bodies and fight corruption through transparent administration of the programme. This legislation came after several years of high economic growth — the ‘India Shining’ years — that, however, failed to

236

Social protection, growth and employment

significantly improve the living conditions of many poor people. In this sense, it may provide a new learning experience for other countries that have failed to translate rapid economic growth into similarly strong poverty reduction. The scale of implementation of this programme grew very fast; job creation accelerated from less than 1 billion workdays distributed among 20 million households in 2006–2007, the first year of operation, to 2.5 billion workdays among 50 million households in 2010–2011. Any initiative of such breadth and ambition is sure to face enormous implementation challenges that can only be addressed through a process of learning and adaptation. There are reports, by both critics and supporters, citing instances of resources being diverted to the pockets of local elites; of partial payments to beneficiaries, below the number of hours effectively worked; of women’s wages paid to their husbands; and of inadequate projects due to a lack of maintenance or simply not being finished, with the result that there is no impact on agricultural productivity. It is interesting and encouraging that many of the ‘failures’ are reported in the programme’s own auditing process. The fact that information on these failures is publicly available will most likely contribute to improving the programme’s design and implementation and can be interpreted as a good signal in the interest of addressing its complexity. The Act has clearly generated a number of very positive results. There is evidence of effective job creation (documented both by the programme’s own audits and by independent studies); wages are being paid; females are actively participating in the programme and are receiving wage payments directly (rather than through their husbands or other male family members); projects undertaken by the programme are considered useful by their communities; general wage levels have increased; and living conditions have improved. A recent India-wide household survey offers evidence that the programme is in fact reaching and benefiting poor people, even though it has not yet been able to effectively guarantee jobs. Unfortunately, there has been no evaluation instrument proportional to the programme’s magnitude that could provide a reliable picture of just how generally successful (or not) it has been. This study seeks to shed light on issues that have not been addressed thus far by previous evaluations. It assumes an effective implementation of the programme under the NREGA — i.e. that the programme is effectively creating jobs for poor people in rural areas, that workers are being paid the official programme wage and that the nation’s castes and tribes are being employed in proportions similar to those stated in the programme’s official figures. With these assumptions, we address the macroeconomic and distributional implications of running an employment generation programme such as the NREGA, including the indirect employment effects it has through its secondary effects on other sectors. Second, we look at the programme’s impact on prices and hence on the cost of living of rural households. Third, we consider the programme’s economic and distributional effects when land productivity increases. Finally, we briefly discuss the extent to which leakages, in the form of hiring non-poor workers, would change the programme’s economy-wide impact, as well as the impact of changing the size of NREGA through a contraction/expansion of its budget.

India

237

Overall, the modelling exercise reported in this chapter indicates that, since its inception, the programme has had a positive impact on economic activity beyond the immediate and direct impact of wage payments to the poor people who participate in it. It finds both that the economy as a whole benefits from the programme and also that each major population group does as well. Poor workers in rural areas benefit the most through direct employment creation. Secondary welfare impacts through the creation of demand in other sectors of the economy are larger for higher-income groups than for poor people. In that sense, the programme’s distributional effects are not positive. The impacts are significant yet small due to the programme’s relatively small size in relation to the Indian labour markets. This negative distributional impact, however, is very small and does not modify the programme’s overall progressive redistribution of income in the Indian economy.

India: Growth, poverty and inequality India has the second largest population in the world, about 1.2 billion people. It is also an emerging economic power, with a strong growth record over the last 30 years. In terms of size, India is the tenth or fourth largest economy in the world, depending on whether market exchange rates or purchasing power parity rates are used to translate local currency into US dollars. Either way, India’s economy clearly stands out among emerging and neighbouring economies. Yet, at market exchange rates, India’s income per capita is only US$1330, and the country has just migrated from ‘low-income’ to ‘low-middle-income’ status.

Economic growth During the first three decades after independence in 1947, the Indian economy grew at the modest rate of 3.5 percent per year which, in the face of a rapidly growing population (at 2 percent per year), was insufficient to significantly reduce poverty. The 1980s, however, marked a turnaround in economic conditions: economic growth accelerated to a rate of 5.5 percent per year, producing an increase in per capita income of about 3.5 percent per year, which opened the door for significant poverty reduction. The removal of widespread government controls on trade and industry in the 1990s and the long-lasting and rapid expansion of international trade resulted in a further acceleration of growth. Most recently the economy has proven resilient to a variety of shocks; droughts, high international oil prices and the global recession did not prevent growth at above 8 percent between 2009 and 2011. The performance of the Indian economy, along with those of other emerging countries, slowed in 2011–2012, and the outlook is now not so bright. The growth forecast for the 2012–2013 fiscal year is 6.5 percent. As the pace of growth has accelerated over the last 30 years, the nation’s economic structure has become increasingly service-based. In 1960, services accounted for about 30 percent of gross domestic product (GDP), while in 2010 they constituted almost 60 percent. India is now widely recognized as a strong world competitor in skills-based services such as information technology. And while the importance of industry has also increased, from less than 20 percent

238

Social protection, growth and employment

to about 30 percent, the share of agriculture has plummeted from slightly more than 50 percent in 1960 to less than 20 percent in 2010. The shift from agriculture to industry and services is the traditional route accompanying economic development. In India, however, the decline of agriculture has been aggravated by a lack of investment and negative incentives that retarded productivity growth in agriculture at the cost of lagging living standards in rural areas.

Poverty and inequality Poverty in India has been falling since 1983 at a varying pace. While poverty reduction was fast during the years that preceded the passage of the NREGA, the country’s subsequent success in reducing poverty did not match its increasing economic growth. Between 1973–1974 and 1987–1988, GDP almost doubled, and poverty decreased by 30 percent, but between 1987–1988 and 2004–2005 poverty decreased again by about 30 percent, while GDP almost tripled. According to government figures, poverty continued to decrease between 2004–2005 and 2009–2010 at a faster pace.2 Unfortunately, as in many other countries, faster growth was also accompanied by rising inequality. India’s Gini coefficient, a widely used measure of inequality, increased from 0.286 in 1993–1994 to 0.305 in 2004–2005 in rural areas and from 0.343 to 0.375 in urban areas. India’s rising inequality also saw the widening of regional disparities across states. Between the 1990s and the 2000s, the variation in income per capita across states increased substantially to accentuate differences in the living conditions of people based on where they lived.3 By the mid-2000s, large differences in well-being across states were evident. The incidence of poverty in different states ranged from 3 percent of the total population in the richer states to 57 percent in some of the poorest. The four states with the lowest poverty incidence housed just 2 percent of all poor people in India, whereas the 14 poorest states accounted for 80 percent. In two states, Orissa and Bihar, more than half the population is poor. Social disparity in India goes back to a history of discrimination against particular population groups. According to the Indian Constitution, the population is classified into four groups: the scheduled tribes (STs), the scheduled castes (SCs), other backward classes (OBCs), and other population groups (others). At about the time of the passage of the NREGA, these groups accounted for 8, 20, 41 and 31 percent of the total population, respectively. Groups that are discriminated against have been and remain at the bottom of the social and economic pyramid and thus include large concentrations of poor people. In rural areas, about 49, 40 and 30 percent of the ST, SC and OBC households, respectively, had monthly per capita consumption expenditures lower than INR410.4 In all other population groups, only 20 percent of households had an income below this level. Deprived population groups also have limited access to land, most notably in the case of the SCs; when the Act became law, about three quarters of ST households had land possessions smaller than 0.4 hectare, and only 1 percent had more than 4 hectares. By contrast, 6 percent of ‘others’ had more than 4 hectares, and 3 and 4 percent, respectively, of SCs and OBCs had that much.5

India

239

Employment conditions and policies Living standards critically depend on employment opportunities. Open unemployment rates are low in India, at 2.5 percent in rural areas and 5.3 percent in urban areas. Low unemployment rates confirm the well-known dictum that poor people cannot afford to be unemployed; they further suggest that low labour force participation might also be a critical issue. India’s employment rate—i.e. the proportion of the population that actually works—is indeed low. It is low among men (at a rate of 55 percent regardless of the area of residence) and particularly low among women at 33 and 17 percent in rural and urban areas, respectively.6 Having a job or a small plot of land does not guarantee freedom from poverty. Wages are low in many rural areas, particularly for women. As an extreme case, the mean wage of an illiterate female worker in rural areas is less than one tenth of the mean wage of a male worker with a university education living in a city.7 In a country where the majority of people still work in agriculture, low agricultural productivity is an important determinant of poverty. Rural workers either work as employees in farms for a low wage or squeeze a living out of small and often low-quality landholdings. Employment has been prominent in Indian policy discourse but less so in development plans. During the initial decades of development planning, the pursuit of growth was to be partially based on labour-intensive small enterprises, whose growth was expected to improve employment conditions. The small-enterprise sector was encouraged by reserving the production of certain goods to this sector and by providing fiscal concessions. Aside from these policies, which ultimately met with little success (Little et al., 1989), employment generation was effectively seen as a byproduct of policies promoting growth and the resulting changes in the structure of production.8 In the face of persistently poor living conditions in rural areas, the government added to the small toolbox of employment-minded policies the design and implementation of employment generation programmes, with a focus on low-income groups. The Seventh Five-Year Plan (1985–1990) and Ninth Five-Year Plan (1997–2002) clearly identified productive employment generation as a major objective, but the overall policy approach to job creation continued to see it as a largely residual effect of growth. Job creation policies supplementing the plan’s emphasis on growth included the promotion of labour-intensive sectors and two major employment creation programmes, the Sampoorna Grameen Rozgar Yojana and the National Food for Work Programme. Compared with the NREGA, these two programmes were quite limited.9

240

Social protection, growth and employment

The Mahatma Gandhi National Rural Employment Guarantee Act The Act The 2005 NREGA guarantees a minimum of 100 days a year of paid employment to all rural households. Building on India’s long-term experience with employment programmes, it adopted a self-targeting mechanism to reach poor people. The works sponsored by the programme are limited to manual, low-skilled tasks remunerated at the state’s minimum wage. Thus, the universal right that the Act establishes reaches poor people through self-selection. Those not interested in working at the minimum wage do not participate. Although the Act does not explicitly indicate a time of year when work should take place, the expectation is that job opportunities will be offered during the agricultural lean season when distressed migration and forced sales of valuable assets are highest due to the lack of local employment opportunities. The flexible employment generation projects offered can thus provide timely support to sustain poor people during this lean season, helping them to avoid actions that might undermine their capacity to sustain living conditions in the future and/or incur burdensome debts. Such projects can also help reduce the economic and social costs of migration. For the work guarantee to be effective, the programme must ensure that jobs are made available such that poor people find them attractive. On the edge of survival, the costs of taking a job need to be considered carefully. In the absence of child-care facilities, women with children might not be able to take the guaranteed jobs even if the additional income is much needed. The Act stipulates that, when needed, work sites must have child-care facilities. It also calls for work sites to provide drinking water, shady places for resting, and first aid kits. Without such amenities, even poor people might not find the jobs attractive. The Act also specifies that work must be provided within a 5km radius of villages; if work takes place at sites that are further away, transportation cost must be added to the workers’ wages. In addition to job creation, the Act also seeks to increase productivity in agriculture, improve environmental management and facilitate access to markets. It specifies that work made available must fall within the categories of: • water conservation and water harvesting; • drought proofing, including plantation and afforestation; • irrigation canals, including micro and minor irrigation works; • flood control and levees; • land development; and • rural connectivity.

India

241

Seeking to ensure that projects are labour-intensive, the Act stipulates that they must allocate at least 60 percent of their total budget to wages and forbids the use of contractors. The Act is not designed to be a rural development programme and does not contemplate building complex development projects. Social inclusion and gender equality rank high among the Act’s objectives; a minimum of one third of the jobs created should be made available to women. It also aims to reach deprived groups, such as the SCs and STs; although it does not specify a target rate for the participation of these groups, it promotes wide dissemination of information and transparency of implementation and requires that the participation of traditionally excluded groups in the programme be regularly reported. Furthermore, it specifies that minor irrigation, horticulture and land development projects can be undertaken on the lands of deprived population groups.10 India’s history highlights the importance of empowering poor people and controlling corruption for poverty alleviation to succeed. The Act and subsequent operational guidelines make elaborate provisions to give poor people control over decisions regarding public works carried out under the Act. A key provision places the selection and monitoring of works within the communities, the gram sabha and panchayat.11 Key anti-corruption and empowering provisions involve job cards needed to qualify for programme benefits; these are granted when a person can verify their name, address and age. With a job card, rural workers can make a submission for the number of days of work of their choice. Because they allow tracking of submissions for work, the number of days worked and the wages received, the cards also provide relevant data on the implementation of the programme. To avoid forgery, the Act specifies that the job card must stay with the worker and that the programme administration must keep a copy. In addition, the programme promotes transparency by mandating that ‘muster rolls’, records of work undertaken and wages paid, are made publicly available at the work site. The Act provides for periodic social audits and the setting up of vigilance committees within the villages. The costs of the programmes sponsored by the Act are shared between the federal and state governments. The central government finances the entire wage payroll of the unskilled workers, 75 percent of the materials costs and 75 percent of the wage bill of skilled workers. State governments cover the remaining 25 percent of the materials costs and the skilled workers’ wage bill and also 100 percent of unemployment allowance if they fail to provide the requested job (Government of India, 2005: 10–11).

Evolution of the programme The Act was inaugurated on 2 February 2006. In its first phase it covered the 200 least developed rural districts, and in the second it covered 330 districts. All rural districts were reached during the third phase of implementation (Table 6.1). During the first year of operation the programme created 1 billion person-days of work that benefited 21 million households. By 2009–2010 this had more than doubled: 2.6 billion person-days of work to 55 million households. The average number of days of work provided rose from 43 person-days per household in its initial year to 50

242

Social protection, growth and employment

person-days in 2009–2010. Its size and scope are unprecedented in the history of India’s social programmes. In the last two years, the programme has decreased in size. It will be important to ascertain whether this is indeed a new trend or simply a temporary slump and in either case to identify the reasons behind it.

Table 6.1: NREGA’s Job Creation and Household Coverage, 2006–2011 Year

Districts (number)

Households Annual Work Average Average (millions) Person-days Annual Days Daily Wage (billions) Per Household (Rupees–INR) (number)

2006/07

200

21.0

0.9

2007/08

330

33.9

1.4

2008/09

615

45.1

2.2

2009/10

615

52.5

2.8

2010/11

615

55.0

2.6

100

2011/12

615

49.9

2.1

118

Source: author’s elaboration based on data from NREGA official data (www.nrega.nic.in).

The participation of women in the NREGA’s programme has been remarkable. From the beginning, female involvement has been well above the minimum prescribed quota of one third, and it has increased over time, from 41 percent of all workdays in 2006–2007 to 48 percent by 2008–2009, since when it has remained about constant (Table 6.2). In contrast, the programme’s delivery on its inclusion commitments to deprived population groups is less encouraging. Although the participation of SC households has increased for the most part, that of the ST households started at a high proportion but decreased during the first three years of implementation. Note also the large decline in the share represented by SCs in 2011–2012. It is beyond the scope of this study to try to explain the reasons for these changes.

Table 6.2: Share in the Total Number of Person-days Worked on NREGA Projects (percentage) by Women, SCs and STs, 2006/07–2011/12 Year

Women

SCs

STs

2006/07

40.6

25.4

36.4

2007/08

42.5

27.4

29.3

2008/09

47.9

29.3

25.4

2009/10

48.1

30.5

20.7

2010/11

47.7

30.6

20.9

2011/12

48.2

22.0

18.2

Source: author’s elaboration based on data from NREGA official data (www.nrega.nic.in).

India

243

It has been argued that an Act that gives a right to work and tries to promote social inclusion should aim to pay the same wage everywhere. In parallel with the implementation of the NREGA, the national government has embarked on the equalization of minimum wages across states in the country. In 2006 the minimum daily wage across states ranged from INR40 to INR90; by 2010 it ranged between INR120 and INR140—a significant reduction in the spread. Although equalizing wages across states with large differences in cost of living is not necessarily an optimal strategy, some reduction in the range is likely to make a positive contribution.12 To increase productivity in agriculture and improve environmental management, the Act has concentrated its efforts on water and land management projects. Since its implementation, these have between them represented more than two thirds of the projects officially reported as completed; the importance of land management has increased from 20 to 30 percent (Table 6.3). Prominent among water management projects are those oriented towards conservation, harvesting and irrigation.

Table 6.3: Distribution of NREGA Works by Project, 2006/07 and 2011/12 Category

Type

2006/07 number

Water management

Land management Connectivity and others

2011/12 %

number

121,921

31.8

232,809

15.6

Irrigation facility

27,362

7.1

207,941

Renovation of traditional water bodies

25,472

6.6

88,420

13.9 5.9

Microirrigation works

12,151

3.2

76,621

Land development

43,370

11.3

234,767

Drought proofing

30,989

8.1

117,785

Flood control and Protection

10,206

2.7

87,238

Rural connectivity

91,244

23.8

374,481

Other activity approved by MRD

20,776

5.4

75,911

Water conservation and harvesting

5.1 15.7 7.9 5.8 25.0 5.1

Source: author’s elaboration based on data from NREGA official data (www.nrega.nic.in).

The budget available to fund the Act’s programmes has increased as employment creation expanded, but the programme’s ‘burden’ on the economy—measured by the size of its budget as a proportion of GDP — only increased in the first three years to a peak of 0.6 percent in 2009–2010 (Table 6.4) then declined due to the programme’s slower growth and the country’s continued rapid economic growth. If the current slowdown in economic growth continues, it is possible that the programme’s burden will rise back to its earlier 2008–2009 level.13

244

Social protection, growth and employment

Table 6.4: Central Government Expenditures on NREGA Programmes, 2006/7–2011/12 Fiscal year

Percentage of GDP

INR crore

NREGA budget estimate

NREGA budget release

NREGA budget estimate

NREGA budget release

2006–07

0.26

0.20

11,300

8,694

2007–08

0.24

0.25

12,000

12,661

2008–09

0.28

0.53

16,000

30,000

2009–10

0.61

0.52

39,100

33,539

2010–11

0.52

0.47

40,100

35,841

2011–12

0.45

0.33

40,100

29,215

Notes: GDP at market prices (current prices with base year 2004–05); data for 2009–10 are provisional, for 2010–11 are based on quick estimates and for 2011–12 are based on advance estimates. Source: Reserve Bank of India (www.rbi.org.in). NREGA expenditure data are sourced from the Ministry of Rural Development (www.rural.nic.in). Source: author’s elaboration based on data from Reserve Bank of India and Ministry of Rural Development.

The impact of the programme Since there are no systematic, nationally representative and independent data on the programme, there are as yet no comprehensive evaluations of it. However, evidence from the programme’s administrative records, independent studies by academic institutions, the Act’s social audits and the 2009–2010 National Household Survey suggest that it is having a significant impact on the lives of poor people. Overall programme performance according to the various sources varies significantly according to the specific feature under scrutiny and the geographical location. Independent studies and social audits detail weaknesses and failures, which, although in many cases they raise concerns about the implementation of the programme, also suggest that on the whole it is lowering poverty. The 2009–2010 National Household Survey data confirm that the programme is reaching poor people and is contributing to social inclusion by increasing the participation of deprived social groups and women; it also confirms a strong variance across states in the degree to which the programme accomplishes its objectives.14 Independent studies and social audits suggest that income in villages has increased since the programme’s inception, with the increases varying from very small to significant—around 20 percent of annual income. As expected, studies report that additional income has been used substantially for food consumption but also to cover education and health expenses and to repay household debt. Of particular importance, the rise in income earned locally has curbed distress migration; estimates of this effect vary widely though, with some finding no visible reduction, while others find a complete elimination of distressed migration.15 The high participation rate of women in the NREGA is perhaps one of its most important achievements. Since women’s participation in other types of paid labour is particularly low,

India

245

a 50 percent average participation in the programmes sponsored by the Act is a remarkable achievement. Similarly important is the provision to equalize male and female wages. As more women join the paid labour market under significantly improved wage conditions, household income and women’s status in the household and the community improves. Studies and social audits confirm these trends but also suggest that the journey toward women’s empowerment will be a long one. Some studies report that women have gained power, have made the decision to work on their own and have improved their livelihood choices, but others have found that many women still hand their wages to their husbands or that husbands make arrangements to receive directly the wages earned by their wives.16 As important as a 50 percent women’s labour force participation rate might be, there are still obstacles preventing women from joining the programme. The decision to take a job under the Act, as any other job, can have costs, which may become prohibitively high for poor people. Particularly relevant for women is the availability of child-care facilities; according to most studies, although the programme has been relatively successful in providing drinking water and shade, it has dome much less well in providing health and child-care facilities for working women.17 This is an area where much needs to be done to improve the programme’s benefits for women. An important aspect of the poverty reduction impact of the NREGA is the increase in the wages received by poor workers. Three important wage effects have followed the implementation of the Act: • the degree to which the wage paid by the Act has raised the local market wage for casual labour in agriculture; • whether the Act is abiding to its requirement to pay the minimum wage, a point further reinforced by the increase in minimum wages; and • to the extent that the guaranteed jobs are effectively paying the legal minimum wage, the Act’s wages have become the de facto wage floor—i.e. poor workers not working under the Act’s programme also benefit from its higher wage provision. This is an area where evidence is more difficult to obtain. Studies report that the Act has increased the wages workers receive when working for the programme, and this is a very important effect. It does not necessarily mean, however, that workers are effectively receiving the stipulated minimum wage. According to studies, the Act does not always abide by the requirement to pay the state’s minimum wage to its workers. Other studies and evidence from the 2009–2010 National Household Survey suggest that employment creation by the programmes associated with the Act are not meeting the demand for jobs in the rural communities, although local wages have indeed increased since the Act’s inception.18 Therefore, even if not fully guaranteeing employment, the Act is most likely helping to improve the lives of poor people even when workers do not directly engage in the programme. The programme’s ability to sustainably reduce poverty largely rests on the adequacy and quality of the assets created. Building relevant and quality rural assets ensures a stronger and longlasting impact on agricultural productivity, which would increase poor people’s consumption

246

Social protection, growth and employment

of food and raise their living standards. The programme’s rapid expansion and sheer size has surely exerted pressures on local managerial capacities to coordinate asset-building projects, so one should not be surprised to learn that asset quality is in need of improvement. Official figures state that in 2010–2011 only 50 percent of projects initiated under the NREGA during the previous fiscal year were actually concluded, suggesting that even by this crude indicator there is need for improvement. Village studies also indicate that asset creation requires more attention. Studies indicate that assets sponsored under the Act are often built with a short-term perspective, do not last long, are of low quality and are not properly maintained. Other studies also indicate instances in which assets are considered valuable by villagers, improve crop yields and have long-term positive environmental effects. Some of the blame for the low quality of assets goes to the specification that projects must involve extensive use of manual work and that 60 percent of the programme expenditures should go directly to labour costs. But other observers disagree, arguing that much can still be done to improve asset quality within the stipulated criteria. Dreze and Khera (2009), for example, suggest that asset quality can be enhanced by a modest use of science and technology coupled with participatory planning so that the right assets are selected and the proper technologies adopted.19

Assessing the NREGA’s economy-wide impact Given the potentially far-reaching effects of employment generation programmes, economywide modelling is a particularly useful tool to analyse their impact in the economy at large. The section below presents the results obtained from an economy-wide model that replicates the characteristics of the NREGA. We identify the beneficiaries according to those defined by the programme, using the coverage and actual composition of workers identified by the programme in 2009–2011, and assuming that the programme only hires low-skilled workers. We also run a separate simulation to look at the impact of an increase in agricultural productivity associated with the implementation of the Act. The data used in this model correspond to a national Social Accounting Matrix (SAM) built for 2003, before the inception of the Act.20 Overall, we find that the employment programme in India has a positive impact on economic activity, as well as direct and indirect positive impacts on the income of poor households. The potential increase in agricultural productivity is found to have a positive economic effect with shared benefits to all population groups, even though its distributional impact is slightly negative to the extent that the welfare of higher income groups increases more than the increase in welfare for poor people.

Assessing the NREGA’s macro-economic impact We first model an employment generation programme with a budget equivalent to 0.65 percent of GDP, which approximately corresponds to the actual size of the NREGA in fiscal year 2009–2010. The simulation consists of an increase in public expenditures equivalent to

India

247

0.65 percent of GDP to hire workers and pay for the necessary materials for construction and the wages of the administrative and few technical staff required by the projects. The simulated increase in public expenditures closely follows the actual composition of the budget in 2009–2010. The payment of wages to workers under the Act is equivalent to 0.43 percent of GDP, expenditures in intermediate inputs are equal to 0.19 percent of GDP, and the payment for government services is equivalent to 0.03 percent of GDP. In proportional terms, 66 percent of programme expenditures go directly to wages for beneficiaries, 5 percent to administration expenses and 29 percent to purchase inputs for the implementation of projects. In the model we draw labour from rural households in proportions approximating those in the Act in 2009–2010. We assume that the labour hired is divided in equal parts between illiterate male and female workers and that these workers belong to the poorest 60 percent of rural households of SCs and STs and to the poorest 30 percent of households of OBCs and others. The payment of wages to these workers is further inputted into the income of the households from which workers are drawn. The cost of intermediate input materials and administrative expenses, which are assumed to include payments to school- and college-educated workers, represent about one third of the budget, in accordance with the actual budget reported by the programme in the period under consideration. Results indicate that running an employment generation programme such as the one under the NREGA has a positive macroeconomic impact. An allocation of resources equivalent to 0.65 percent of GDP increases GDP by about 0.4 percent. The programme’s overall expansionary effect is in accordance with the basic notion of a Keynesian balanced budget multiplier, plus the additional demand generated from the shift in income towards poor households with a high marginal propensity to consume. The programme’s distributional impact is also positive. Simulation results indicate an increase in welfare among poor rural households and a marginal increase among poor urban households. The implementation of the programme carries a cost, which comes in the form of a decline in welfare for rich people in both urban and rural areas, since the programme is financed through income taxes and ‘forced’ savings. Overall the Act generates an expansion of activity and changes in the composition of production across sectors that result in a progressive redistribution of income towards poor people. The main effects are a sizable redistribution from rich people in urban and rural areas to poor people in rural areas, a marginal redistribution from the same groups to poor people in urban areas and an overall redistribution from urban to rural areas. These results are generated by a variety of factors, some of which are discussed in the sections below.

Shift in consumption As the NREGA is implemented and wages are paid to workers, the demand for goods and services consumed by poor people in rural areas rises. Simultaneously, the increase in taxes to finance the projects sponsored by the Act reduces the demand for goods consumed by rich urban

248

Social protection, growth and employment

households. Thus, the demand for goods and services shifts towards those consumed by lowincome households in rural areas. This, in turn, induces changes that boost prices and economic activity in agriculture and light manufacturing and reduce the demand for other manufacturing and service sectors. At a greater level of detail, economic activity increases in the production of rice, the other processed food sector, textiles and apparel; conversely, it decreases in vehicle manufacturing and in most services.21 The pattern of changes in income suggests that employment generation programmes such as the NREGA are an effective instrument for poverty reduction. Poor rural households benefit from employment opportunities directly through the programme as well as from the employment opportunities generated indirectly by the implementation of the programme. Through indirect channels the benefits from the programme extend well beyond its direct beneficiaries. In our simulations, these positive effects extend to 90 percent of the rural population and 30 percent of the urban population. Although the effects are small, partly due to the small size of the programme relative to the size of the Indian labour market, they are nonetheless noticeable.

Impact on the factors of production According to our results, the implementation of the NREGA increases labour income and income from land but decreases income from capital. The largest absolute income variation is the change in labour income; we focus on these changes below. The Act’s impact on the economy triggers a wide array of changes in labour income, varying by area, sex, education and caste/tribe.22 The implementation of the programme increases the demand for workers with basic skills over the demand for workers with higher skills. Reinforcing the programme’s objective of benefiting unskilled rural workers, the largest indirect increase in labour income occurs among illiterate rural workers, with a smaller increase for rural schooleducated workers. Female workers fare slightly better than males, and the income of workers that belong to the ST increases more than for any other group. In contrast, urban workers and rural workers with higher education are negatively affected. Within these groups, the labour income of females falls more than male income. When one considers caste/tribe, income of the SCs tends to fall the most. Overall, the Act has a positive impact on labour income and has a progressive impact on the distribution of wages. To the extent that changes in the income of different groups of workers move in different directions, it is useful to aggregate them according to labour characteristics. Aggregation shows that changes in labour income favour workers in rural over urban areas, less-educated workers gain more than better-educated workers, females workers fare better than males, and, although on a smaller scale, workers in deprived population groups fare better than all others. It is worth noting at this point that the increase in labour income triggered by the programme — i.e. the indirect rise in labour income—is in the order of 0.11 percent. The small size of this effect is mainly because the direct wage payments of the programme represent no more than 8 percent of the total income earned by illiterate workers and 2 percent of the total annual income of the

India

249

rural labour force. This result serves as a reminder that, as impressive as the NREGA programme is, it is still a small fraction of labour income in India.

Price impacts Implementation of the NREGA is likely to prompt changes in consumption prices, which affect the cost of living of various population groups. According to our results, the Act increases the prices of most goods and reduces the prices of services. As income shifts to poor people, and their demand for goods increases, the prices of the goods they consume also go up. Therefore, parallel to the increase in the income of poor households, there is also an increase in the cost of living which counters the Actâ&#x20AC;&#x2122;s initial welfare impact. The implied increase in the cost of living, however, is small. Poor rural households that do not benefit from the expansion of employment under the NREGA will be negatively affected by an increase in their cost of living. Albeit very small, this is still an issue that would require further thinking about ways to compensate, if needed, such negative effect.

The economic and distributional effects of rising agricultural productivity Next we consider the results of the second core simulation exercise: the NREGAâ&#x20AC;&#x2122;s potential increase in land productivity. The simulation consists of an exogenous 1.5 percent increase in the factor productivity of land. The exercise assumes that the pre-existing pattern of land ownership remains in place and that productivity increases uniformly across all lands. Increasing the productivity of land has the expected effect of expanding economic activity. The simulated 1.5 percent increase in land productivity produces an appreciable rise in GDP, final demand and trade, ranging between 0.02 and 0.03 percent. The increase in agricultural productivity has a less positive impact on income distribution. Although the rise in productivity increases welfare across all household groups, the increase is larger for rich people and larger in urban than in rural areas. The differences are considerable, particularly in rural areas. In cities, rich people gain three to five times more than poor people; but in rural settings, rich people gain 10 to 20 times more than poor people. Even if undesirable, these results are to be expected, mainly due to the high concentration of land in rural India.

The impact on the economy The increase in productivity means that the same quantity of agricultural products can be generated with fewer quantities of labour, land, capital and inputs from other sectors. But as the production in agriculture expands, more labour is used, along with more capital and land. The increase in factor use is due to the downward effect of higher productivity on factor prices, which allows producers to buy more labour, capital and land. A more productive agricultural sector has

250

Social protection, growth and employment

an expansionary effect on other sectors, although the changes are smaller. Accordingly, the use of more labour and capital extends to sectors such as food manufacturing, textiles and services. Not all sectors experience an expansion in activity; the change in relative prices leads to less activity in heavy manufacturing and extractive activities. On the whole, production shifts toward sectors that tend to hire low-skill workers from poor, rural households. It should be noted that the labour impact of the increase in productivity has the opposite effect of the one resulting from employment generation features of the Act. The expansionary effect of the rise in land productivity is strong enough to increase the income accruing to all three factors of production. Proportionally, the change is greatest for labour, then for capital and finally for land. The productivity hike significantly pushes down the relative prices of agricultural commodities and manufactured food. Because food represents a large proportion of poor householdsâ&#x20AC;&#x2122; consumption basket, the change in commodity prices reduces their cost of living. Our results also indicate that the relative prices of services and heavy manufacturing goods increase.23 And given that these figure more visibly in rich householdsâ&#x20AC;&#x2122; consumption basket, the cost of living for these households goes up. The effect on prices is so strong that even if the productivity hike decreases poor peopleâ&#x20AC;&#x2122;s income and, hence, consumption expenditures, the fall in the price of food implies such a strong rise in purchasing power that they end up better off than before. After the increase in productivity, poor people can buy more food and perhaps even increase purchases of other goods. It is important to note that the difference between the two measures is wider for poor people in rural areas and rich people in urban areas, suggesting that these two groups might experience stronger changes in purchasing power than others. At the macroeconomic level, the positive effect of improving land productivity on GDP opens the opportunity to reduce income tax rates and still maintain the budget in a balanced position. If, instead of reducing income taxes, the government decides to reduce value-added taxes, it can reduce the negative distributive effects that follow the rise in productivity while keeping the budget in balance. Because value-added taxes place a disproportionally higher toll on poorer people, low-income households will benefit more than high-income households. Although the changes in taxes are small, over time they may provide useful financing for complementary policies seeking to reduce poverty. The potential increase in agricultural productivity is found to have a positive economic effect with shared benefits to all population groups, even though its distributional impact is slightly negative to the extent that the welfare of higher income groups increases more than the increase in welfare for poor households. Our simulation increasing land productivity has the expected result of expanding economic activity and enhancing aggregate welfare. It increases welfare and consumption expenditures across all households, both rural and particularly urban, and it amplifies activity in the agriculture, food-processing and service sectors. To the extent that the welfare and consumption of all

India

251

households improve, poverty is reduced. By expanding economic activity, particularly by reducing the cost of food, the increase in land productivity greatly contributes to the improvement of living standards. This suggests that increasing productivity should be a main objective when making decisions about what type of works should be financed through the NREGA. Projects that increase land productivity generate the largest welfare benefits for poor people, even if they are not directly employed by the programme. In the simulation results, however, increasing productivity also has a negative distributional impact due to a larger increase in the welfare of rich households. This result is driven by the high concentration of land in the hands of rich people. This result should not be interpreted as a criticism of the important objective of programmes such as the NREGA to increase productivity in agriculture. Instead, our results reinforce the importance of targeting investments on the landholdings of poor households, which can potentially bring larger welfare benefits.24 The simulation results conform to the well-known effects of the green revolution, which greatly reduced the cost of producing food and made food reliably available to vast sectors of the population, thereby contributing to the reduction of poverty. Accordingly, many studies have identified the reduction in food prices as its most important contribution. Similarly, studies have also pointed to its depressing effect on agricultural wages. Our simulation clearly shows that the latter outweighs the former, resulting in a reduction of poverty.

The effect of leakage Poverty reduction programmes run the risk of missing poor people and attracting population groups different from the intended target. Our analysis suggests that the effects of leaks do not visibly change the programmeâ&#x20AC;&#x2122;s macroeconomic impact. The major difference between a programme with and without leaks lies in the distributional impact. But even here, the change is small. Beyond the obvious difference in income flows resulting from spreading the same amount of resources over a larger population, the distributional impact is not very different. Allowing for leaks has no significant implications for how the programme affects sectors, factor markets and commodity prices. The results analysed above thus suggest that the presence of leakage in the implementation of a programme such as the NREGA should not be a major concern for policy makers considering its macroeconomic and distributive effects.

Changing the size of the NREGA Testing for the effects of expanding or contracting the size of the programme or modifying the assumed increase in land productivity suggests that size changes do not qualitatively affect the impact of the programme. Reducing or increasing the employment programmeâ&#x20AC;&#x2122;s budget by 50 percent mainly results in an amplification or deflation of the impact on macroeconomic variables and its distributional effects. The same effect is obtained from varying the 1.5 percent increase in land productivity to 1 or 2 percent. This suggests that policy makers do not need to

252

Social protection, growth and employment

worry about the possibility that further expansions of the programme could trigger deleterious economic consequences. Likewise, no qualitative gain should be expected from reducing the size of the programme.

Conclusion This study has shown that the NREGA’s employment programme has positive economic effects. This suggests that its immediate poverty reduction effect takes place at no cost to the economy more broadly or to the country’s GDP in particular. The Act generates a virtuous redistributive effect, since the programme benefits accruing to poor people are not limited to the wages directly paid by jobs created by the programme. Through the changes the programme generates in the economy, the implementation of the Act results in further job creation in rural and urban areas –jobs that are likely to be taken by poor people — reinforcing the benefits to poor people in rural areas and extending them to those in urban areas. The increase in labour income generated by the new jobs created by the programme is small, but the small size of the induced changes is largely due to the still comparatively small size of the programme. Furthermore, the cost of the programme is manageable. In our simulations, while government expenditures for the programme are funded by an increase in income taxes, only paid by rich people in urban areas, its implementation ends up imposing, at most, small reductions in consumption and welfare among rich people in both urban and rural areas. The study also finds a potential cost of concern to poor people. As the economy expands and poor people improve their lot, the price of the commodities they consume is likely to increase. This increase in poor households’ cost of living is small and should not be a matter of concern, as the programme’s benefits more than compensate for the increase in prices. However, for poor rural residents that do not benefit from the programme’s jobs, even this small rise in the cost of living would be important. This suggests the need to complement the Act with programmes aiming to provide support to poor non-beneficiary households. The programme’s economic impact and indirect distributive effects are not significantly affected if it fails to reach poor people in rural areas at 100 percent efficacy. Even if the size of the programme has no precedent in India or elsewhere, our findings show that the programme has positive macroeconomic and distributive effects that are proportional to its budget. It thus suggests that although it represents a significant administrative challenge, the rapid expansion of past and future enlargements do not carry negative macroeconomic or distributive effects, at least within the range of changes analysed. If a reasonable increase in land productivity is assumed after the implementation of the Act, the short-term benefits of the Act’s employment creation are coupled by long-term benefits that decrease poverty and enhance the economy. However, the benefits of higher land productivity do not accrue to poor people. This is not an argument against the Act, because it is significantly determined by the strong concentration of land. This finding suggests the need to reinforce the Act’s provisions aiming to improve land productivity on the lands of poor populations. The Act’s employment creation and land productivity actions complement each other well, as in the case of

India

253

offsetting effects on the price of food, rendering the Act a potentially powerful tool to decrease poverty in the immediate and long term.

Notes 1 “An Act to provide for the enhancement of livelihood security of the households in rural areas of the country by providing at least one hundred days of guaranteed wage employment in every financial year to every household whose adult members volunteer to do unskilled manual work and for matters connected therewith or incidental thereto.” National Rural Employment Guarantee Act, 2005; no. 4, 2O F2 005, 5 September 2005. 2 Poverty estimates and hence the change in poverty have been the subject of intense debate in India. The cited figures only intend to help illustrate the point. 3 The coefficient of variation of state income per capita increased from about 0.33 during the first half of the 1990s to about 0.35 in the second half of the same decade and further to about 0.39 in 2004–2005. 4 As reference, India’s poverty line in 2006 has been estimated to be INR447 and INR579 per month in rural and urban areas, respectively (Government of India, 2009). 5 These figures come from NSSO (2006). Figures for population by social group come from p. 21, for land possessions from p. 25, and for consumption from p. 27. These figures understate inequalities, as land possession does not say anything about the quality of possessed land. 6 Note that this employment rate has the total population as reference and not only the working-age population. Data correspond to 2004–2005 and come from NSSO (2006: 35). The employed population refers to the usually employed population – i.e. those who worked for the longer part of the 365 days preceding the survey or worked a minimum of 30 days during the reference period of 365 days preceding the survey (SIC): “The workforce according to the usual status (ps + ss) includes persons who (1) either worked for a relatively longer part of the 365 days preceding the date of survey and (2) also those persons from among the remaining population who had worked at least for 30 days during the reference period of 365 days preceding the date of survey.” (p. 34). The 2009–1010 NSSO survey portrays a similar picture (note, however, that the corresponding data, table S8, p. 35, refers now to the population aged 15 to 59 years). 7 Figures are the average wage received by regular waged/salaried employees. In 2004–2005 the ratio of male urban to female rural wage was 2.4 and remained the same in 2009–2010 (for casual workers, the ratio was 1.9 in this later year). See NSSO (2006: 92) and Government of India (2011: 92–96). 8 See Papola (2006). 9 Three other employment programmes have been started or revamped after the inception of the NREGA: the rural self-employment programme, Swarnajayanti Gram Swarozgar Yojana; the urban self-employment and salaried employment programme, Swarna Jyanti Shaharri Rozgar Yojana; and the subsidized credit programme to create employment in rural and urban areas, the Prime Minister’s Employment Generation Programme. See Government of India (2010: 22–25). 10 Mainly on the lands of SCs and STs, of households registered as below the poverty line, of deprived households whose needs have already been recognized by making them beneficiaries of the Indra Awas

254

Social protection, growth and employment Yojana housing programme and land reform and, more generally, on the lands of small and marginal farmers who constitute the majority of the holdings in most states.

11 Gram sabhas and panchayats are village meetings and institutions that constitute the local self-government of villages and small towns – grams – with at least 500 people of voting age. 12 The coefficient of variation halved between these two years, from about 20 to 10. 13 Recall that the central government accounts for most of the cost of the programme but not for the entire bill. 14 See, for example, Dutta, Murgai, Ravallion and van de Walle (2012). 15 See, for example, Jandu (2008), Panda, Dutta and Prusty (2009), Indian School of Women’s Studies Development (2008), Hirway and Singh (2006) and Singh and Nauriyal (2009). 16 See, for example, Kamath, Murthy and Sastry (2008) and Dreze and Khera (2008). 17 See, for example, Singh and Nauriyal (2009) and Pankaj and Sharma (2008: 190–210). 18 See, for example, Dreze and Khera (2008), Centre for Food and Agribusiness Management (2009) and Kamath, Murthy and Sastry (2008). 19 Dreze and Khera (2009: 12) argue that the economic return of the assets that the NREGA builds might not be very different from the return on other investments, including those of many industrial projects. 20 We use the SAM-based economy-wide STAGE model to probe the potential economic impact of an employment programme such as the NREGA. We implement the model in its comparative, static mode. We choose not to use the recursive, dynamic mode because of three varied considerations. First, the time path of the implementation of the Act’s programme was not a subject for investigation, given that it was set by the Act; second, the development of a business-as-usual baseline for the recursive dynamics requires the imposition of a large number of additional assumptions; and third, the magnitude of the effects of the Act’s programme are sufficiently small that they will be dwarfed by the underlying growth and factor productivity changes in the Indian economy. Hence we focus on the Act’s ‘short-term’ distributional implications, the key policy objective, rather than on the ‘longer-term’ growth implications, a secondary policy objective. For a full description of the methodology, see Zepeda et al. (2013). 21 For a full description of the multiple impact of the NREGA on the economy through the interrelation of economic activity and induced demand for labour, see Zepeda et al. (2013). 22 Note that these changes in labour income do not include the jobs created and wages paid directly by the NREGA. 23 The model’s numéraire is the Consumer Price Index (CPI). As one set of relative prices rise, so must another set decline; however, the CPI weights are drawn from the average consumption bundle, and hence the implications vary by household according to the consumption bundle of households. 24 As such, the alleviation of relative poverty through such targeting would be via the second ‘law’ of welfare economics effects.

India

255

References Centre for Food and Agribusiness Management, Quick Appraisal of 5 districts under NREGS in Uttar Pradesh, Report submitted to Ministry of Rural Development, Government of India, Indian Institute of Management, Lucknow, 2009. Dreze, J. and R. Khera, ‘The Battle for Employment Guarantees’, Hindu Frontline, 2008, 26(1). Dutta, P., R. Murgai, M. Ravallion and D. van de Walle, ‘Does India’s Employment Guarantee Scheme Guarantee Employment?’, Policy Research Working Paper, No. 6003, World Bank, Washington, DC, 2012. Government of India, ‘The National Rural Employment Act, 2005’, Gazette of India, Ministry of Law and Justice, New Delhi, 2005: 10–11. Government of India, Review The Methodology For Estimation Of Poverty, Planning Commission, New Delhi, 2009. Government of India, Annual Report to the People on Employment, Ministry of Labour and Employment, New Delhi, July 2010: 22–25. Hirway, I. and H. Singh, Concurrent Monitoring of NREGA: Feedback From the Field, Ministry of Rural Development, New Delhi, 2006, www.tinyurl.com/hirway-singh (accessed 19 Dec. 2012). Indian School of Women’s Studies Development, Impact of NREGA on the Living and Working Conditions of Women in Rural India, Ministry of Rural Development, New Delhi, 2008, www.tinyurl.com/iswsd-nrega (accessed 19 Dec. 2012). Jandu, N., A Study on Socio-Economic Empowerment of Women Under NREGA, Ministry of Rural Development, New Delhi, 2008, www.tinyurl.com/jandu-2008 (accessed 19 Dec. 2012). Kamath, R., R. Murthy and T. Sastry, NREGA Surveys in Anantapur, Adilabad, Raichur and Gulbarga (2007–08), Ministry of Rural Development, New Delhi, 2008, www.tinyurl.com/kamath-et-al-2008 (accessed 19 Dec. 2012). National Sample Survey Organization (NSSO), Employment and Unemployment Situation Among Social Groups in India 2004–05, NSS 61st Round, report no. 516, NSSO, New Delhi, 2006: 35. Panda, B., A.K. Dutta and S. Prusty, An Appraisal of NREGA in the States of Meghalaya and Sikkim. Professional Institutional Network, Ministry of Rural Development, New Delhi, 2009, www.tinyurl.com/panda-et-al-2009 (accessed 19 Dec. 2012).

256

Social protection, growth and employment

Pankaj, A. and A. Sharma, ‘Issues in Implementation and Impact of NREGS: Bihar and Jharkhand Experiences’, The National Rural Employment Guarantee Act (NREGA): Design, Process and Impact, UNDP, New Delhi, 2008: 190–210. Papola, T.S., Employment Trends in India, Institute for Studies in Industrial Development, New Delhi, 2006, http://isid.org.in/pdf/EmployTrenz.PDF (accessed 19 Dec. 2012). Singh, S.P. and D.K. Nauriyal, System and Process Review and Impact Assessment of NREGS in the State of Uttarkhand, Ministry of Rural Development, New Delhi, 2009, www. tinyurl.com/singh-nauriyal (accessed 19 Dec. 2012). Zepeda, E., S. MacDonald, M. Panda and G. Kumar, Employing India: Guaranteeing Jobs for the Rural Poor, Carnegie Endowment for International Peace and UNDP, New York, 2013.

7 The Youth Employment Challenge: The insertion of Kenyaâ&#x20AC;&#x2122;s youth into formal and informal jobs Eduardo Zepeda, Fatou Leigh, Lydia Ndirangu and Jacob Omolo Introduction Kenya faces a significant unemployment problem that affects young people especially hard. Youth unemployment rates are several times higher than adult rates and particularly high in cities and among females. As young people grow up, they break the link with their family and hopefully become economically independent. During the transition from childhood to adulthood, access to good jobs of acceptable quality is essential for young people if they are to gain independence from their parents, improve their job market opportunities and enhance their prospects of forming a family. During the last four decades, and in recognition of Kenyaâ&#x20AC;&#x2122;s unemployment and underemployment difficulties, successive government administrations have assumed and prioritized employment creation as a core policy (Republic of Kenya, 1969; 1983; 2008a; 2008b). More recently, policies aiming at employment creation have also focused on young people. Important initiatives in this regard are the Sector Plan for Labour, Youth and Human Resource Development Sector (2008â&#x20AC;&#x201C;2012) and the KaziKwaVijana (KKV) programme. Launched in April 2009, the KKV aims at annually employing between 200,000 and 300,000 young people in rural and urban areas in labour-intensive public works projects that are implemented by different government ministries (Office of the Prime Minister, 2011).

258

Social protection, growth and employment

This chapter uses the 2005/2006 Kenya Integrated Household Budget Survey (KIHBS) to describe the main characteristics of Kenya’s youth employment challenge in search of clues to inform youth employment policies. It adopts a working definition of youth as those aged from 15 to 34 years, which mirrors the most comprehensive official Kenyan definition, but also highlights the different age subgroups that might be worth distinguishing in the context of youth employment policies. The first section of the chapter considers the main dimensions of unemployment and their conditioning factors– gender, education, income and area of residence. The second examines youth labour earnings, concentrating on people reporting wage income. The third offers a brief review of Kenya’s employment and youth employment policies. The chapter concludes with some summary remarks.

The youth employment challenge In 2005/2006, because of the large youth bulge due to fast population growth, people of 15 to 34 years of age comprised two thirds of the population of working age (15–64),and many of them face the hardships of unemployment. While population growth has been receding, it is still high, and the number of people aged 15 to 34 years will continue for several decades to grow faster than the adult working-age population (Figure 7.1). Youth pressures on labour markets do appear to have begun to lessen, which may eventually make it easier to reduce youth unemployment. Kenya’s population dynamics underscore the importance of further reducing fertility rates through measures such as family planning, as suggested by Kenya’s Human Development Report 2010. Consistent with the normal pattern, youth unemployment is considerably higher than that of any other age group (Figure 7.2); the national unemployment rate is about 10 percent, while rates for people between 15 and 25 years reach a multiple of this, with the peak occurring around 20 years of age. Understanding youth employment challenges necessarily involves looking at the entire range of activities in which they may engage. As young people reach the age at which school is no longer the only choice, they have to opt between staying there or engaging in homebased activities or joining the labour market in pursuit of a paid job. As they shift from school to the search for jobs, it is almost inevitable to find high unemployment rates, partly due to the time the search takes. We distinguish six main activities that people engage in: those who work in a formal job, work in the informal sector or are searching for a job make up the ‘economically active’ group; those who choose unpaid domestic work, full-time study or other activities are the ‘economically inactive’. Formal employment is here defined as paid employment in registered private companies, in the public sector (including central and local government, Teachers Service Commission (TSC), state-owned and government-controlled enterprises) and in non-governmental or international organizations. Informal employment refers to working employers, own-account

Kenya

259

Figure 7.1: Population Trends by Age Group and Area of Residence, 1955–2100 Annual (millions of people)

Age-composition (percentage shares)

Percent share

1 0

Millions

Youth Adult

1950

1960

1970

1980

1990

2000

2010

Year

70 80 Percent

Rural Urban Share of rural

Millions 20 30

Rural/Urban population (millions and percentage)

1960

1980

2000

2020

2040

Year

Source: authors’ construction based on UN Population estimates. Note: data points show five-year age cohorts in a smoothed form.

workers, unpaid family members, apprentices, other workers and paid employees working for an individual or an undefined (other) employer.1 As the share of young people in full-time studies declines rapidly between 15 and 25 years of age, work in the informal sector becomes the dominant activity, although unemployment also rises steadily from 3 to 23 percent between 15 and 20 years of age, as young people struggle to get a job, after which it decreases continuously to about 10 percent at the age of 35and a bit further during adulthood (Figure 7.3). Early school drop-out puts pressure on the youth labour market. By the ages of 30 to 34 years the composition of young people’s activities has become quite similar to that of the rest of the adult population; the big transitions have been left behind. The proportion of young people in formal jobs only begins to become visible around the age of 25; by the age of 35 it has increased to about 15 percent. At this preliminary level, employment

260

Social protection, growth and employment

promotion policies might emphasize insertion for young people aged 15 to 28 years and promotion for those between 29 and 34 years.2

Figure 7.2: Youth and Adult Unemployment Rates, 2005/06 (percentage)

20 10

Percent

Smooth Actual

Age

Source: authors’ calculations based on KIHBS (2005/06) micro records.

60 40 0

Percent share

100

Figure 7.3: Activity Choices of the Working-age Population 2005/06 (percentage)

Years of age

Source: authors’ calculations based on KIHBS (2005/06) micro records.

Kenya

261

Youth unemployment is significant in both urban and rural areas but is greater in urban areas, particularly among the youngest. In rural areas unemployment ranges between 20 and 25 percent for those aged 15 to 25 years, while in urban areas it ranges from 35 to 60 percent (Figure 7.4). The difference between rural and urban unemployment decreases as age increases; high urban unemployment is partially the result of rural-to-urban migration that tends to flood the urban job market. The proportion of young people living in rural areas decreases from about 90 percent at the age of 15–18 years to 70 percent at the age of 30–34 years (Zepeda et al., 2012). Thus the lack of attractive opportunities in rural areas re-emerges as an urban youth unemployment problem.3 The urban setting is known for the variety of opportunities it offers; in particular the proportion of young people with formal jobs is systematically higher in urban than in rural areas (Figure 7.4). Rural labour markets, in turn, are clearly dominated by the early insertion of young people in informal employment activities, most prominently traditional farming. More people, including young people, engage in home-making and other miscellaneous activities (neither formal nor informal). Thus unemployment, either as a rate or as a proportion of the corresponding age bracket, is higher in urban areas; relatively fewer people take on informal activities, while more search for formal employment but often do not find it. Unemployment is significantly higher among young females than among young males—by more than 10 percentage points for those aged 15 to 25 years and by a smaller amount at higher ages. At its peak female unemployment approaches 50 percent, compared to a male rate of slightly over 30 percent; at around 34 years of age, the female rate is a little above 15 percent, while the male rate is just above 5 percent. Females have fewer job opportunities, and at each age a smaller share of females have a job in both the formal and the informal sectors (Figure 7.5). The females’ relative disadvantage in accessing informal jobs is worst at around 25 years of age. In the case of access to formal jobs, that disadvantage widens throughout the entire youth period. The employment problems of young females persist despite their substantially greater involvement than males in home-making activities and other miscellaneous chores. The proportion of young females who are home-makers ranges from 6 percent in the early years to 14 percent at around 30 years of age (Figure 7.5, lower centre panel). Such involvement reduces the pressure females exert on labour markets, but if their interest in other work continues to increase, there will be more female pressures on labour markets.4 Part of the stronger pressure exerted by young females on labour markets originates in the lower proportion than males that attend school on a full-time basis (Figure 7.5, lower left-hand panel), a gap of 10 percentage points throughout most of the period of youth. The combination of high unemployment and low school attendance calls for careful consideration in employment promotion policies: poorly designed policies might induce increases in female employment at the expense of postponing, perhaps indefinitely, their education. This reinforces the need to design employment promotion policies with a gender focus. Unemployment is very high for young people with primary and secondary education, in the latter case reaching a peak of over 35percent for those around 20 years of age (Figure 7.6). Unemployment is even higher for those with no formal education, but they represent only a very

262

Social protection, growth and employment

Figure 7.4: Youth Unemployment Rates and Activity Choice by Area, 2005/06 (percentage)

Unemployment rate, percent

Rural Urban

50 40 30 20 10 0 20

10 20 30 40

20 40 60 15

35 15

2 4 6 8

2 4 6 8 10

Share in population

0 20 40 60 80 0 10 20 30

Age

Source: authors’ calculations based on KIHBS (2005/06). Note: data plotted in a smoothed form. Unless otherwise specified all line plots represent data smoothed by a cubic spline smoother with the degree of smoothing automatically set through cross-validation by the software.

Kenya

263

Figure 7.5: Youth Unemployment Rates and Activity Choice by Gender, 2005/06 (percentage)

Male Female

Unemployment rate, percent

0 20

35 15

30 25 20 15 5 10 20

35 15

2 4 6 8

5 10 15

Share in population

0 20 40 60 80 0 5 10

10 20 30 40 50 60 70

Age

Source: authors’ calculations based on KIHBS 2005/06.

264

Social protection, growth and employment

30 20 0

Unemployment

Figure 7.6:Youth Unemployment Rates and Activity Choice by Educational Attainment, 2005/06 (percentage)

20 30 10 0

0 15

35 15

10 15

0 5 10 15 20 25

Share in Population

0 20 40 60 80 100 0 20

20 40 60

Age

Source: authors’ calculations based on KIHBS 2005/06.

Kenya

265

40 0

Unemployment

Figure 7.7: Youth Unemployment Rates and Activity Choice by Gender, Area of Residence and Age, 2005/06 (percentage)

35 15

40 30 0 10 20 20

35 15

0 5 10 15

Share in Population

0 20 40 60 80 0

10 20

20 40 60 80

Age

Source: authors’ calculations based on KIHBS 2005/06.

266

Social protection, growth and employment

small share of the population. At age 25â&#x20AC;&#x201C;30, unemployment exceeds 30percent for those with tertiary education; although the number of people involved is small compared to those with primary and secondary education, this indicates just how general the problem of unemployment is. Access to a formal job increases with education; the unemployment rate by level of education echoes the access young people have to jobs.5 Very few young people with no or just primary education have a formal job (Figure 7.6, top left-hand panel); for those with secondary education it reaches around 25percent by age 35, and for those with tertiary education it gets to be over 80percent by that age. As new cohorts with more education enter the labour force, each of these percentages is likely to fall. Meanwhile, educational level is in general inversely related to having an informal job (Figure 7.6, top centre panel). Young people with secondary education are proportionately less devoted to domestic work and other miscellaneous activities than those with primary education. Despite significantly better eventual access to formal jobs than those with only a primary education, their unemployment figures are relatively high (Figure 7.6), presumably associated with greater reluctance to take informal jobs and greater time spent searching for a job. This is the typical pattern found in developing countries. Residence, gender, education and income do not condition choice of activity in isolation from each other. A look at gender in conjunction with area of residence, for example, reveals that unemployment affects urban females the most and rural males the least.

Figure 7.8: Unemployment Rate by Gender, Type of Area of Residence and Education, 2005/06 (percentage)

Percent

No education Primary Secondary

Source: authorsâ&#x20AC;&#x2122; calculations based on KIHBS 2005/06.

Kenya

267

Within each gender, youth unemployment is usually highest for those with no formal education. Unemployment rates of young people with secondary education are different according to area and gender: male and female rates in rural areas increase between the ages of 15 and 25 years; rates for urban males decrease continuously from 15 to 34 years, but rates for urban females only start to decrease for age groups older than 22 years (Figure 7.7).6 Thus, increasing education tends to lower unemployment rates, particularly among young workers older than 25 years. These unemployment patterns relate to young people’s access to formal and informal jobs. Particularly, the high unemployment of urban females with primary and secondary education appears to be related to their distinctively low access to informal jobs (Figure 7.8).The better access that education grants to formal employment is an advantage enjoyed by urban males with a tertiary level but is much less apparent for other groups of young people.7 In summary, youth unemployment is a major challenge in both rural and urban areas, in the former because of the large numbers involved and in the latter also because of the very high rates. Significant differences by gender, area of residence and age group (within the broad official definition) suggest that policies need to focus on finer age subgroups than is currently the case. Although urban unemployment rates are significantly higher than those in rural areas,

Figure 7.9: Employment and Unemployment by Age, Type of Area of Residence, Gender and Education, 2005/06 (percentage of the population) 20

Share in population

Female rural Female urban Male rural Male urban

Age

Source: authors’ calculations based on KIHBS 2005/06.

268

Social protection, growth and employment

employment promotion policies should also look carefully at rural areas. One important reason for this is that the number of unemployed young people is high in rural settings, but there are at least two other important reasons. First, the lack of job opportunities in rural areas surfaces in part as urban unemployment (as young people migrate to cities), meaning that urban unemployment levels could be reduced by successful employment promotion in rural areas. Second, the young inactive population is proportionally larger in rural areas, which means that a sudden availability of jobs might prompt them to switch from economic inactivity to participation. More rural jobs might â&#x20AC;&#x2DC;paradoxicallyâ&#x20AC;&#x2122; result in more unemployment, as discouraged workers might decide to join the labour force, but this would on balance be a positive development. Unemployment plagues young people from a very young age up to about 28 years, depending on their education, gender and other social factors. Quantitatively the unemployment problem is greatest for young people between 18 and 24 years. Female unemployment is worse than that of males. Young women in rural areas account for the largest number of unemployed people, and females in urban areas have the highest rates of unemployment. Youth employment promotion policies must aim to provide jobs to females in rural and urban areas, perhaps by establishing minimum quotas. They also need to consider that females might face a higher opportunity cost when attempting to work, as they may also have home-making duties. Thus youth employment promotion policies should give proper consideration to the cost of accessing jobs. Although young people with no formal education have some of the highest unemployment rates, the most pressing problem is that of those with primary and secondary education. Youth employment promotion policies should not, therefore, restrict themselves to those with only basic skills but, rather, give priority to generating jobs for people with primary and secondary education. Given that unemployment among tertiary-educated people is not high and is limited to certain age groups, actions to improve the functioning of labour markets for people with tertiary education might be the most appropriate form of intervention. Access to jobs and household income are directly interwoven, with unemployment rates consistently higher for young people from the bottom 40 percent of the income pyramid. Pressure from high-income households to aim promotion policies towards their constituency can be great, given that the number of unemployed people from these households is also high at certain ages. Appropriate policies to address their employment needs might be of the active type, whereas more decisive interventions might be needed to address the unemployment problems of young people from low-income households. Targeting poor people should be an important component of employment promotion policies. Unemployment is, of course, related to the availability of informal and formal jobs. Since the data seem to suggest that the market is setting a significant premium on secondary education, not to mention tertiary education, youth employment promotion policies should avoid introducing disincentives to continue school and complete secondary education: policies might in fact consider introducing incentives to combine work and school. Given the need to improve apprenticeship programmes, policymakers might consider apprenticeships with a combination of vocational training and formal education.

Kenya

269

Youth unemployment is due directly to the scarcity of formal job openings and the incapacity of young people to create an attractive enough job in the informal market for themselves or to find one working under someone else. Youth employment policies should thus emphasize the promotion of micro, small or household businesses, both for young people and more generally, to widen opportunities to work as owners or employees of these businesses. Promoting young females in these businesses should also be emphasized. Finally, the overall promotion effort should ensure that, wherever possible, policies benefit first people from low-income households.

Making a living: labour income and young people The still strong presence of traditional rural activities in the Kenyan economy is reflected in the fact that not all employment generates monetary income. This underscores the importance of distinguishing between work that is performed under traditional or non-monetary arrangements and work that takes place on a monetized basis. According to the 2005/06 KIHBS, there were 4.9 million people earning a monetary income from their labour (either a salary or the earnings from a business). Such employment is of lesser importance in rural areas but predominant in urban settings; about 20 percent of the rural population fell into this category, compared to 45 percent in cities. In terms of the number of people, the sheer size of the rural population means that the majority of people that earned some income from work are found there (2.9 million), compared to 1.9 million people in urban areas.8 Men are more likely to earn money from their labour than are women, particularly in rural areas. Of those receiving some income for their work, the most important category is employees. There are 1.9 million paid employees in rural areas receiving a payment and 1.4 million in urban areas, although this category constitutes a higher share of total employment in urban areas. Second in importance among those receiving a monetary income are some 850,000 own-account workers, about 580,000 of them in rural areas and 270,000 in urban areas. Other categories constitute only a small number of people.9 We estimate wages as the total labour income reported by paid employees and include in our calculations only those paid employees actually receiving a monetary payment. Wages in Kenya, as in many other countries, are distinctively higher in urban than in rural areas.10 While average rural wages typically vary by age from KShs3,000 to KShs9,000 per month, in urban areas they are between KShs5,000 and KShs25,000 (Figure 7.10; note the different scales). The age pattern of wages varies with area of residence. As of 2005/06 the average urban worker reached an earnings peak somewhere between 40 and 50 years of age, while the average rural worker did so at past 50. The pace of the increase in mean wages throughout younger years is almost constant in rural areas, whereas in urban areas it starts slowly and then accelerates around the age of 30 (Figure 7.10). Wages depend heavily on the type of employer. They are highest in the formal private sector and lowest in informal businesses (Figure 7.11) and higher in cities than in rural areas. The urban wage advantage is larger for older age groups, especially among employees in the private

270

Social protection, growth and employment

sector and, to a lesser extent, among other informal workers, although in this case we should talk of income from labour and not of wages. On the other hand, the upward trend with age is less marked among employees in informal businesses and hardly visible among public-sector employees. Finally, as expected, the wage floor for paid work corresponds to the mean wage of employees in informal businesses, which exhibits small differences either by area of residence.

Figure 7.10: Wage Variation by Age in Rural and Urban Areas, 2005/06

KSh., thousands

Employee Other

Age

Source: authors’ calculations based on KIHBS 2005/06.

Figure 7.11:Wages by Type of Employer and Location, 2005 /06 Rural

Urban

KSh., thousands

15 20 25 30 35 15 20 25 30 35

Age

Source: authors’ calculations based on KIHBS 2005/06.

Kenya

271

Figure 7.12: Wages by Gender, 2005/06 Male

Female

KSh., thousands

15 20 25 30 35 15 20 25 30 35

Age

Source: authors’ calculations based on KIHBS 2005/06.

Male wages are generally higher than those of females. Among young people, this is clearly the case in informal jobs and in the public sector (Figure 7.12). However, the opposite is the case for young workers aged 15–24 years in the public sector and for young people of all ages in the private formal sector. These last two results are at first glance surprising but are due to the differing composition of male and female employment by level of education and type of employer. To better understand the wage differences between males and females, we sacrifice age detail to gain information on the type of employment. Thus we look at the wages of five age groups in 10 employment categories. Among young people, the mean wage paid to females is higher than that paid to males in non-governmental organizations and government-controlled enterprises, slightly higher in central government and practically the same in the private formal sector (Figure 7.13, left-hand panel—the bars on the left show a higher wage for young women than for young men). Female wages are lower for the rest of the employers — namely, TSC, state-owned enterprises, individual (informal) employers, local government and international organizations. The wage advantage females have in adulthood is significantly different. Females have a large advantage over males in international organizations, a small edge in state-owned enterprises and central government, almost the same wage in local government but a smaller wage in the other types of employers (Figure 7.13).11 Of course, some of these differences are due to varying levels of education, as will be shown below. As expected, average wages increase with formal education, most notably for those with tertiary education (Figure 7.14). For employees of private formal businesses, where the contrast is sharpest, the wage of a 25-year-old with tertiary education is three times that of someone with secondary education, and at around 34 years the wage difference is five-fold. Though smaller, this gap in public-sector and informal businesses is also significant—around double by age 30 and higher for older groups.12

272

Social protection, growth and employment

Figure 7.13: Wages by Type of Employment for Young People and Adults by Gender, 2005/06

NGOs

Individual

Other

T.S.C.

Private

Local gov.

Central gov.

State-owned enterprise

International organization

Gov. control enterprise

Male Female

KSh., thousands International organization

Individual

Local gov.

Other

State-owned enterprise

T.S.C.

Private

Central gov.

Gov. control enterprise

NGOs

KSh., thousands

Male Female

Source: authors’ calculations based on KIHBS 2005/06.

Figure 7.14: Wages by Education, Age and Type of Work, 2005/06

KSh., thousands

No education Primary Secondary Tertiary

20 25

30 35 15 20 25 30 35 15 20

25 30 35 15 20

25 30

Age

Source: authors’ calculations based on KIHBS 2005/06.

There are also significant wage differentials among young people with secondary, primary and no formal education, but both the levels and the percentage gaps are much smaller. The differential between secondary and primary reaches three to one (Figure 7.14) in the private sector by age 34 and is twice as high in the public sector; in the informal work sector, it is much less. On the other hand, there is not much difference between the wages of a young person with no formal education and one with primary education, except in the public sector; this may be partially explained by the wage structure of public bureaucracy.

Kenya

273

It appears that, for young people, education really makes a difference in the wages they can hope to earn, but also that the interaction between age and type of job is very important. Young people with primary education see their wages improve when they are older if they happen to have a job in the public sector. Young people with secondary education can hope to earn a higher wage when they are older in any sector, more so in the public and private formal sectors but also in the informal sector. A statistical exercise analysing wages by age, education, type of area of residence, gender and type of employment (among other variables) confirms the analysis presented here (Zepeda et al., 2012). It suggests that if everything else is equal, males earn significantly higher wages than females, and urban workers have higher wages than rural workers. The exercise also confirms that education and access to a formal job are decisive factors in determining wages. In particular, the exercise shows that while age is an important determinant of wages, education is much more important. This suggests that workers with secondary and, especially, tertiary education are in strong demand in the economy. The idea that there is a strong demand for young people with secondary education is consistent with the rapid reduction in unemployment among this group of people as age rises. This suggests that any youth employment promotion scheme should not interfere with the pursuit of secondary education; better, policies promoting youth employment should provide incentives for young people to pursue and acquire secondary education. The wage premium of tertiary education is so large that it goes without saying that the supply might well be falling short of the marketâ&#x20AC;&#x2122;s needs. In summary, people working for pay represent a significant proportion of the employed Kenyan population. Large numbers of people work for pay in urban areas, where they represent the majority of workers, but the largest numbers are to be found in rural areas. The majority of these jobs are for employees in informal businesses, with males dominating the work-for-pay sphere. Young people tend to work in informal businesses, as adults tend to capture a higher share of the jobs in the formal private sector and, particularly, in the public sector. This suggests that youth employment policies should consider targeting an increase in job opportunities for those aged 15â&#x20AC;&#x201C;24 years in the private and public sector; this could perhaps be done through wide apprenticeship programmes. Wage structures confirm that jobs in the public and formal private sectors are the most attractive occupations. Wages by type of employer suggest that the lowest wage paid to formal employees is the mean wage paid in informal businesses, slightly higher in cities than in rural areas and slightly higher for older young people. The analysis also confirms that there are significant wage differences between males and females. Any programme aiming to create jobs should look closely at the lowest wage paid in formal jobs and determine to what extent such a wage represents a socially satisfactory reward for work in public work programmes and to what extent a higher wage might inhibit otherwise economically viable activities. It should also pay close attention to measures that could narrow wage differences between males and females. The large wage premium for workers with tertiary education and the opportunities for wage improvement opened by secondary education underscore the need to increase education

274

Social protection, growth and employment

in Kenya. Any youth employment promotion programme should make its initiatives compatible with the aim of increasing education. This study did not look at the impact of apprenticeships or vocational training on earnings. However, it highlighted that the size of current programmes pale in comparison to the need for bringing more young people into apprenticeships. It is recommended that any scaling up of apprenticeships should not be done in isolation from efforts to increase education.

Employment and youth policies in Kenya Kenya has a long record of implementing employment policies. Over 40 years ago, for example, the 1970–74 Development Plan (Republic of Kenya, 1969) featured employment and unemployment as important policy matters. Later, in the mid-1980s, the Report of the Presidential Committee on Unemployment (1982/83) and the Sessional Paper No. 2 of 1985 on Unemployment were important components of employment policy development. More recently, The Sector Plan for Labour, Youth and Human Resource Development Sector (2008–2012) looks into unemployment as a structural issue, explicitly adding the youth dimension. In 2009the government launched the youth-focused KKV programme, aiming to create 200,000 to 300,000 jobs in labourintensive public works. In 2010, this programme was further extended to include training and internships, with a strong focus on infrastructure projects. This section briefly reviews Kenya’s employment policy record, with special attention to the main recent initiatives addressing youth employment issues. It also quickly reviews some international employment generation policies. Over the years, Kenya has experimented with a broad set of employment policies. Between 1963 and 2011, more than 17 types of policies have been implemented at one point or another, with employment (or more broadly labour market outcomes) as one objective. There has been considerable continuity, since 10 of the 17 policy areas have been a constant in the menu throughout the entire period. To facilitate review, we group the measures into three types—structural policies, active labour policies, and macroeconomic or horizontal policies — and distinguish three periods: 1963 to 1979, 1980 to 1989, and 1990 to the present (Table 7.1). Macroeconomic management can be considered the predominant policy from 1990 to 2011; active labour market policies, with a certain emphasis on macroeconomic measures, were the predominant instrument from 1980 to 1989; and structural and Kenyanization policies were the predominant instrument from 1963 to 1979. In each of these periods, policies were augmented with various short-term interventions that aimed to address particular issues of employment generation. On the other hand, interventions such as wage restraint, economic growth, industrial and agricultural promotion, public works programmes and active labour market policies are present in all three periods. Although there has been a range of policy initiatives directed at the employment challenge, the continuously difficult employment situation implies that they have been inadequate. That said, it is difficult to assess whether these policies failed because of their focus and/or the manner

Kenya

275

Table 7.1: Typology of Kenya’s Employment Creation Interventions, 1963–2011 Employment Creation Interventions

Period 1963–1979

1980–1989

1990–2011

√√

√√√

√

√√√

Structural Policies Infrastructure Development Rural Development Kenyanization

√√√

Industrial Policy

√√

√

√√

Informal Sector Development

√√

√

Productivity Promotion

√√

√

Agricultural Promotion

√√

√√√

√

√√√

√√

√√√

√

√√

√√√

√

√√

√

√√

√√√

Labour Policies Public Works Wage Restraint Active Labour Market Policies Tripartite Agreements

√√√

Education and Training Employment and Labour Market Policies Macroeconomic or Horzontal Policies Economic Growth Macroeconomic Management Legal and Legislative Reforms Fiscal Measures

√√ √

√

Source: author’s elaboration based on government documents. Note: The emphasis that these policies have in each of the three periods is represented by the number of the check mark in the corresponding box. Three check marks indicate that the specific initiative is a top policy, two marks indicate a measure that has medium importance, one mark indicate that the policy was complementary; finally, a blank indicates that the particular intervention was not present or had a marginal importance.

in which they were implemented, or because of the contextual factors present at the time. The question is, therefore, whether well-designed and implemented policies could have significantly improved the employment situation, given the combination of strong demographic pressures and faltering economic growth. This remains an open question, since the policies have not been adequately evaluated. The most that can be said at this level of generality is that the Kenyan policy experience calls for the explicit consideration of three aspects: policy focus, scale of the interventions, and adequate learning from evaluations. The last 10 years have seen the implementation of major youth employment policies with an emphasis on public works, entrepreneurial development and skills enhancement. There have also been attempts to formulate policies promoting employment of Kenyans in foreign countries.

276

Social protection, growth and employment

A key Kenyan institution for youth skills development is the National Youth Service (NYS). It was established in 1964 to offer young people training in vocational, technical and professional skills. It also acts as a reserve force for the Kenya Armed Forces and undertakes rehabilitation and training of disadvantaged and orphaned young people (Republic of Kenya, 2008b). The NYS enrols an estimated 3500 servicemen and women per year on artisanal, craftsmanship and other diploma courses which run for a period of one, two and three years, respectively. It also offers training to young people aged between 18 and 22 years. The minimum qualification criterion is a Kenya Certificate of Secondary Education (KCSE) with a mean grade above D+. In addition, trainees must be unmarried and without dependants. NYS recruitment is carried out in 285 district centres, and there are 16 training institutions spread throughout the country with a capacity of 10,000â&#x20AC;&#x201C;15,000 trainees at any given time. (As a point of reference, the number of people entering the Kenyan labour force in a given year is currently about 600,000).13 The NYS is financed through the national budget, complemented by external funding from the governments of Japan, Germany, the Netherlands and the United Kingdom. Although in recent years there has been a small increase in its budget, funding for the NYS has generally been tiny, at less than 1 percent of the national budget.14 In 2009 the Kenyan government launched the nationwide KKV programme, aiming to employ between 200,000 and 300,000 young people annually in rural and urban areas on labourintensive public works projects (Republic of Kenya, 2010) that are implemented by different government ministries. The KKV creates employment by implementing manual-based small projects in communities, with durations of three to six months. In rural areas they include building dams and irrigation, repairing boreholes and access roads, clearing bushes, sowing organic fertilizers and seeds and planting trees. In urban centres, they include building and operating water kiosks, developing and implementing waste management systems and repairing and maintaining access roads. The KKV uses low wage rates as a self-targeting criterion for young people, allowing for task-oriented and daily payments. Daily wages are based on national minimum wage guidelines issued by the Ministry of Labour, while the task rates are pegged at KShs250 per task for the KKV employees in cities and municipalities and KShs150 for those in all other areas.15 The KKV programme is implemented under the overall supervision and guidance of a national steering committee chaired by the Prime Minister and comprising the Ministers and Permanent Secretaries with KKV projects; programme coordination and monitoring are the responsibility of the Office of the Prime Minister (OPM). Most of the jobs created have been under the Ministry of Water and Irrigation, followed by the Ministry of Roads (Figure 7.15). Other jobs were contributed by the Ministry of Environment and Natural Resources and the Ministry of Forestry and Wildlife. The programmeâ&#x20AC;&#x2122;s budget, which is funded through the national budget, saw a significant increase from KShs3.4 billion (US$43 million) in the 2008/2009 fiscal year to KShs6.6 billion (US$84 million) in the 2009/2010 financial year (Republic of Kenya, 2010). To increase the scale and consolidate the gains made under the KKV, in 2010 the Kenyan government launched the Kenya Youth Empowerment Project (KYEP), a US$60 million, fouryear project funded by the World Bank coordinated by the OPM and designed to enhance the

Kenya

277

Figure 7.15: Number of Young People Employed in KKV Public Works by Ministry Water irrigation Roads Forestry, wildlife Environment & natural resources Regional development Local government Youth affairs Planning Prime minister 0

20,000

40,000

60,000

Number employed in public works

Source: official data available at www.kkv.go.ke/ .

KKV by providing internships and youth training and enhancing the capacities of the implementing agency, the Ministry of Youth Affairs and Sports (MOYAS). It focuses on promoting community participation in identifying projects, establishing clearer project selection criteria and optimizing the labour content of the KKV as well as improving target-setting and developing a monitoring and evaluation system. The renewed KKV gives priority to projects in road maintenance, small-scale water supply and sanitation, water harvesting, forestation and waste collection. An internship and training component of the KKV aims to increase the employability of young people by providing skills through internships and training in the private sector. It targets 10,800 young people aged 15–29 years with a minimum of eight years of schooling for those who have been out of a school/work environment for at least one year. Internships are programmed for between four and six months, with the time split evenly between the workplace and with an identified training provider. The interns are placed in key Kenya Vision 2030 growth sectors such as energy, tourism, information and communication technology, manufacturing and micro and small enterprises. Under this arrangement, each intern receives a monthly stipend of KShs6000, while employers receive a monthly reimbursement of KShs3000 per intern. According to the programme, young people aged 15–17 years are only eligible to take internships in the informal sector.16 This component will link with existing interventions such as the Youth Enterprise Development Fund (YEDF) and institutions such as the Kenya National Federation of Jua Kali Associations (KNFJKA). The third component of the KYEP will provide analytical support to the government in key areas relevant to young people. It will also support capacity-building in youth-oriented institutions within government, provide financial studies in

278

Social protection, growth and employment

critical-to-youth policy and labour market issues and promote institutional capacity-building in institutions such as the MOYAS. Apprenticeships can play an important role for young people by smoothing the transition from school to the job market, helping them gain experience and skills and expanding their job opportunities. Despite government efforts to strengthen the practice of apprenticeship, coverage appears very limited in light of the national youth employment landscape. According to the KIHBS, there were only about 30,000 people in apprenticeships at the time of the survey (of which 95 percent were under 35 years). This means that only 0.3 percent of employed young people or 0.15 percent of all young people were doing an apprenticeship. Although some expansion has subsequently occurred, the need to scale up further is apparent. The quick review of youth employment policies in Kenya indicates that recent policy interventions are, in general, in accordance with the accumulated experience of polices aimed at promoting youth employment across developing countries (see Zepeda et al., 2012). It also suggests that policies are in line with the international experience with public works programmes.17 None of these, however, waive the need for continuous evaluation of such policies and discussion among development stakeholders. Four important issues will perhaps need a more thorough discussion: • the proper scale of policy intervention; • a comprehensive review of Kenya’s experience with youth employment promotion, aiming to inform a more integrated approach to tackle the challenge; • an overall improvement of education, including a significant expansion of apprenticeships; and • the fine-tuning of policies to address the most pressing employment needs of young people, some of which have been highlighted in this chapter.

Conclusion Demographic trends have been one factor behind Kenya’s youth employment challenge. The pressure exerted by young people on labour markets has begun to lessen, however, and this may increase the effectiveness of policies aimed at improving labour market opportunities for young people. The profile of Kenyan youth unemployment as portrayed by the 2005/06 KIHBS suggests a need to focus especially on those aged 18 to 25 years; second, those aged 15 to 18 and 25 to 29 years; and third, those aged 30 to 34 years. While youth unemployment rates are certainly higher in urban areas, rural areas should not be overlooked, in part because the number of unemployed young people is larger in rural than urban areas and in part because a well-functioning rural youth employment programme may be expected also to alleviate urban unemployment by diminishing the pressure on young people to migrate from rural to urban areas. Nor does creating rural jobs in any sense go counter to economic progress and development. While rural-to-urban

Kenya

279

migration is part of any normal development process, it may in a given country proceed too fast from an economic efficiency point of view, and this may be the case in Kenya. Tackling rural youth unemployment will not fundamentally change the long-term economic attractiveness of cities, but it can contribute to a more rational and humane rural-to-urban migration and development process. It is worth insisting that youth employment promotion policies need to keep a strong female focus, in light of the very high female youth unemployment rates. Policy designers might consider, for example, including minimum participation quotas for females but should also address situations in which the cost of taking a job is higher for females than males (e.g. because of safety considerations, travelling distance and family care responsibilities) through measures that reduce that cost. Youth unemployment is also clearly linked to poverty, so targeting poor people should be another aspect of employment promotion policies, not so much because of the size of the group but because their overall employment situation is the worst. Given that the lowest wage paid in rural and urban areas is that of employees in informal businesses, policies seeking to create jobs in public works or apprenticeships or by promoting household enterprises must use this wage as a reference point, while also trying to ensure that the jobs contribute meaningfully to family income. When targeting poor people, care must be exercised to control the understandable pressures that young people from high-income households may exert to gear employment promotion policies towards their constituency, as large numbers of them in certain age groups are unemployed. Improving education leads to better employment outcomes, and the evidence suggests that employee jobs in public and private formal businesses may in practice be increasingly restricted to those with at least secondary education. This in turn suggests than unemployment policies should focus on young people with primary and secondary education, as they constitute the great bulk of youthunemployment, and their situation is much more acute than that of those with tertiary education; although unemployment rates of this latter group can be high, they do not persist for too long after entry into the labour market, and the number of people involved is small. Youth employment promotion policies should give highest priority to those with primary and secondary education; then those in their teens without formal education and, finally, those with tertiary education. The age pattern of unemployment and employment outcomes and the structure of earnings indicate that employers prefer people with tertiary and, to a lesser extent, secondary education. Labour outcomes suggest that young people with secondary education have positive employment outcomes as they approach adulthood; this may be evidence of the payoff not only to education but also to experience, in which case employment promotion policies should support people to acquire the skills and capacities that markets appear to be demanding, by expanding programmes that foster work experienceâ&#x20AC;&#x201D;for example, via wider apprenticeship programmes. The evidence reviewed confirms that youth employment promotion policies should aim to increase both education and employment. It may be better in the long-term to get very young

280

Social protection, growth and employment

people back to school to complete primary or secondary education than to provide employment that competes with their education. A mix of incentives towards employment, school attendance and apprenticeship schemes may be in order. The promotion of formal employment is a complex topic. This study limits itself to pointing out that employment promotion policies should design incentives to increase female participation in formal jobs and to promote the hiring of young people in the public sector. Promoting successful informal employment should be an important component of employment policies. High youth unemployment is related to the scarcity of formal job openings and the incapacity of young people to create an attractive job for themselves in the informal market. Formal jobs will inevitably be restricted to a minority of all job seekers for some time into the future, so they cannot provide more than a partial answer to youth unemployment. Meanwhile it takes time for young people with some level of education to start an informal business or activity. Youth employment policies should emphasize the promotion of micro, small or household businesses for young people as well as for adults who may hire and/or mentor those young people. Kenya has accumulated significant experience in tackling youth employment challenges. The government, for example, has recently launched another important youth employment initiative. Are present policies enough or does Kenya needs to do more to tackle youth unemployment? The overall sense from this study is that it does, that analysis and evaluation need to be more fully incorporated into policy making and that a strong focus must be kept on certain subgroups of the youth population. This study suggests that Kenyaâ&#x20AC;&#x2122;s youth unemployment problems are not simply a reflection of the general unemployment challenge that most developing countries face on their path to development but, rather, a particularly severe problem brought on by the combination over several decades of only modest economic expansion with enormous population growth which created a very high ratio of new entrants to existing participants in the labour force. The initially very small formal sector has not increased quickly enough to absorb the majority of these entrants, nor has the informal sector been able to provide adequate incomes to the bulk of them. Accelerating growth will be part of the answer to this challenge, but more will need to be done. As the Kenyan government lets its current policies run their course, evaluation efforts should be applied to better understand the role of education in labour market insertion, the type of specific gender policies that need to be implemented, the apprenticeship programmes that should be escalated, the promotion measures that should be implemented to support young people as they undertake informal activities, and the specific employment creation programmes that need to be put in place to attend the earning and living conditions of poor people in both rural and urban areas. Heavy emphasis should be placed on what factors determine the demand for labour in both formal and informal sectors, to take advantage of such policies as can accelerate the growth of that demand.

Kenya

281

Notes 1 Arguably, some of the workers included in our groups might be better placed in the other category – for example, working owners of registered businesses or in professional services, which are placed in the informal category, are typically classified as formal. However, instead of devoting our effort and readers’ attention to the definition of formal and informal work, we decided on definitions that give us the broad contours of employment and that could be immediately derived from official data. 2 Of course, not all of the young people leaving school join the labour force. Many become homemakers, and others dedicate themselves to miscellaneous tasks. 3 The cost of seeking a job in urban areas must be lower than the ‘cost’ of staying or migrating to rural areas where unemployment is low. This is an issue that is beyond the scope of this study, so we do not attempt to provide any evidence. 4 The difference between the proportion of females and males who do other miscellaneous chores as their main activity is not large, but there is still a slightly larger proportion of females than males aged between 15 and 25 years who do so. 5 The primary, secondary and tertiary education categories include people whose highest educational achievement corresponded to at least one year of the relevant category. 6 To better analyse this issue, we exclude young people with tertiary education from the discussion, since their numbers are small. 7 These patterns emerge in more detail in a parametric analysis of choice of activity in Kenya using the same 2005/06 KIHBS information (Zepeda et al., 2012). Having a secondary or tertiary education reduces the probability of working in an informal job and increases that of working in a formal job. The econometric exercise also confirms that the probability of having a formal job is greater for males than females, and higher in urban than rural areas. 8 According to the figures in the KIHBS, 77 percent of the working-age population lives in rural areas. 9 If apprentices do receive a wage, this is comparable to the wage for a young employee. Incidentally, note that even if the sample of the household survey severely underestimates the size of youth apprenticeships, it might be safe to say that any reasonable correction for underestimation might still result in numbers suggesting the need to expand apprenticeships. 10 The wage and income from labour figures reported here refer to workers that received some monetary remuneration or income from their labour and that worked for a least one hour during the reference week. The group of other informal includes self-employed and working owners, among others, but excludes people working in traditional agriculture or farming and people that worked in a family business not for pay. 11 Some of these differences can be attributed to variations in the mix of education and occupation within each of the four gender-area groups. 12 The category of ‘other informal’ in our data might include professionals, which will explain the importance of tertiary education and also the fact that their presence is only apparent among older youth age groups.

282

Social protection, growth and employment

13 Based on the current estimate of population at around 42 million people, the estimated age structure and assuming that 70 percent of a given age cohort enter the labour force. 14 Until the late 1980s, the NYS had a compulsory pre-university programme. This programme targeted successful graduates of the Kenya Advanced Certificate of Education (KACE) for six-month preuniversity training. The goal was to inculcate in students the right attitude towards work and instil a culture of tolerance and nationalism. This programme was discontinued in 1990, probably due to its limited impact and sustainability challenges. Since then, the NYS training has been voluntary, with a market-driven focus. 15 The wage for young people working as employees in informal business ranges between KShs2000 and KShs3000 per month in rural areas and between KShs3000 and KShs5000 in urban areas. Expressed in daily rates, considering 22 working days per month, wage rates come to KShs90â&#x20AC;&#x201C;140 in rural areas and KShs140â&#x20AC;&#x201C;230 in urban areas. Attention must be paid to ensure the adequacy of wages paid in employment creation programmes. 16 The wage is thus just above the mean rate for employees in informal businesses. 17 For recent reviews of public works programmes, see Devereux and Solomon (2006), Lal et al. (2010), McCord (2009), del Ninno et al. (2009) and Zepeda and Alarcon (2011).

Kenya

283

References Betcherman, G., M. Godfrey, S. Puerto, F.Rother and A. Stavreska, ‘A Review of Interventions to Support Young Workers: Findings of the Youth Employment Inventory’, Social Protection Discussion Paper, No. 0715, World Bank, Washington, DC, 2007. Bunch, N.-H. and D. Verner, ‘Revisiting the Link Between Poverty and Child Labour: The Ghanaian Experience’, Journal of Economic Literature, 2000. Coenjaerts, C.E., C. Ernst, M. Firtuny and D. Rei, Youth Employment, Youth Employment Network, OECD, Paris, 2009. Dantwala, M.L.,‘Search for an Employment-Oriented Growth Strategy: A Discussion’, Economic and Political Weekly, 1990, 25(21): 1147–1162. Dehejia, R.H. and R. Gatti, ‘Child Labor: The Role of Income Variability and Credit Constraints Across Countries’, Revision of NBER Working Paper, No. 9018, National Bureau of Economic Research, Cambridge, MA, 2003. del Ninno, C., K. Subbarao and A. Milazzo, ‘How to Make Public Works Work: A Review of the Experiences’, Social Protection Discussion Paper, No. 905, World Bank, Washington, DC, 2009. Devereux, S. and C.Solomon,‘Employment Creation Programmes: The International Evidence’, Issues in Employment and Poverty Discussion Paper, No. 24, International Labour Organization, Geneva, 2006. Edmonds, E.V. and N. Pavcnik, ‘Child Labour in the Global Economy’, NBER Working Paper, No. 10317, National Bureau of Economic Research, Cambridge, MA, 2004. Godfrey, M.,‘Youth Employment Policy in Developing and Transition Countries–Prevention as well as Cure’, Social Protection Discussion Paper, No. 0320, Social Protection Unit, World Bank, Washington, DC, 2003. Hafkin, N.J., Gender, Information Technology and Development: An Overview, International Labour Organization, Wageningen, 2002. International Labour Organization (ILO), Employment Policy and Programme for Kenya, ILO East Africa Multidisciplinary Advisory Team, Addis Ababa, 1995. International Labour Organization (ILO),‘Key Indicators of the Labour Market’, KILM 2009, 6th Edition, ILO, Geneva, 2009. Kulundu, D.M.,‘Globalization and Labour Market in Kenya’, Discussion Paper, No. 6, KIPRA, Nairobi, 2003. Krueger, A.B.,‘Observations on International Labour Standards and Trade’, NBER Working Paper, No. 5632, National Bureau of Economic Research, Cambridge, MA, 1996.

284

Social protection, growth and employment

Lal, R., S. Miller, M. Lieuw-Kie-Song and D. Kostzer, ‘Public Works and Employment Programmes: Towards a Long-Term Development Approach’, IPC Working Paper, No.66, International Policy Centre for Inclusive Growth, Brasilia, 2010. Macharia, K., ‘Migration in Kenya and its Impact on the Labour Market’, paper prepared for the Conference on African Migration in Comparative Perspective, Johannesburg, 4–7 June 2003. McCord, A., ‘Cash Transfers: affordability and sustainability’, Project Briefing, No. 30, Overseas Development Institute, London, 2009. McCord, A. and Slater, R., Overview of Public Works Programmes in Sub-Saharan Africa, Overseas Development Institute, London, 2009. Ministry of Youth Affairs,’ Kenya National Youth Policy’,Working Paper, No. 66, International Policy Centre for Inclusive Growth, Brasilia, 2006. Okojie, C.E.E., ‘Employment Creation for Youth in Africa: The Gender Dimension’, paper prepared for the Expert Group Meeting on Jobs for Youth: National Strategies for Employment Promotion, Geneva, 15–16 January 2003. Omolo, O.J.,‘An Overview of Incomes and Wage Policy in Kenya: Effectiveness and Constraints to Productivity Enhancement’, paper presented during a Productivity Policy and Legal Framework Workshop, Mombasa, 2002. Omolo, O.J.,‘The Dynamics and Trends of Employment in Kenya’, IEA Research Paper, No. 1/2010, Institute of Economic Affairs, Nairobi, 2010. Omolo, O.J.,‘Labour and Employment Inequalities in the Context of the East African Regional Integration Process’ in Society for International Development, East African Integration: Dynamics of Equity in Trade, Education, Media and Labour, Ascent Limited, Nairobi, 2011. Pollin, R., An Employment Targeted Economic Program for Kenya, Political Economy Research Institute and UNDP, New York, 2008. Puerto, O.S., Meta-analysis of the Youth Employment Inventory, World Bank,Washington, DC, 2007. Republic of Kenya, National Development Plan, 1964–1970, Government Printer, Nairobi, 1964. Republic of Kenya, National Development Plan, 1970–1974, Government Printer, Nairobi, 1969. Republic of Kenya, Sessional Paper No. 10 of 1973: Employment, Government Printer, Nairobi, 1973.

Kenya

285

Republic of Kenya, Report of the Presidential Committee on Unemployment, 1982/83, Government Printer, Nairobi, 1983. Republic of Kenya, Sessional Paper No. 2 of 1985: Unemployment, Government Printer, Nairobi, 1985. Republic of Kenya, National Development Plan, 1994–1996, Government Printer, Nairobi, 1994a. Republic of Kenya, Recovery and Sustainable Development to the Year 2010. Sessional Paper No. 1 of 1994, Government Printer, Nairobi, 1994b. Republic of Kenya, National Development Plan, 1997–2001, Government Printer, Nairobi, 1997a. Republic of Kenya, Industrialization.Sessional Paper No. 2 of 1997, Government Printer, Nairobi, 1997b. Republic of Kenya, National Poverty Eradication Plan, Government Printer, Nairobi, 1999. Republic of Kenya, ‘Effective Management for Sustainable Economic Growth and Poverty Reduction’, National Development Plan, 2002–2008, Government Printer, Nairobi, 2002. Republic of Kenya, Economic Recovery Strategy for Wealth and Employment Creation, 2003–2007, Government Printer, Nairobi, 2003a. Republic of Kenya, Report of 1998/99 Labour Force Survey, Central Bureau of Statistics, Ministry of Planning and National Development, Government Printer, Nairobi, 2003b. Republic of Kenya, Key Gender and Employment Issues in Agriculture and Rural Development As Aspects of Poverty Reduction Strategy: Kenya Country Study, Ministry of Labour and Human Resources Development, Nairobi, 2005. Republic of Kenya, Vision 2030 Strategy for National Transformation: Accelerating Equitable Economic and Social Development for a Prosperous Kenya, Ministry of State for Planning, National Development and Vision 2030, Nairobi, 2007. Republic of Kenya, First Medium-Term Plan: Kenya Vision 2030–A Globally Competitive and Prosperous Kenya, Office of the Prime Minister, Ministry of State for Planning, National Development and Vision 2030,Nairobi, 2008a. Republic of Kenya, Sector Plan for Labour, Youth and Human Resource Development, 2008–2012, Ministry of Labour, Nairobi, 2008b. Republic of Kenya, 2009 Kenya Population and Housing Census, Volume IC, Kenya National Bureau of Statistics, Nairobi, 2010a.

286

Social protection, growth and employment

Republic of Kenya, 2009 Kenya Population and Housing Census, Volume II, Kenya National Bureau of Statistics, Nairobi, 2010b. Republic of Kenya, Economic Survey, Government Printer, Nairobi, 2011a. Republic of Kenya, Second Annual Progress Report on the Implementation of the First MediumTerm Plan (2008–2012), Monitoring and Evaluation Directorate, Ministry of State for Planning, National Development and Vision 2030, Nairobi, 2011b. Semboja, H.H.H., ‘A Concept Paper on Promoting Opportunities for Youth Employment in East Africa’, paper prepared for the ILO Regional Office and presented at the EAC Meeting of Labour Commissioners, Nairobi, December 2005, World Bank, Washington, DC. United Nations Development Programme (UNDP), Kenya National Human Development Report 2009: Youth and Human Development–Tapping the Untapped Resource, UNDP, Nairobi, 2010. Urdal, H., ‘A Clash of Generations? Youth Bulges and Political Violence’, International Studies Quarterly, 2006, 50: 607–629. Ware, H., Pacific Instability and Youth Bulges: The Devil in the Demography and the Economy, Australian Population Association, Canberra, 2004. Zepeda, E. and D. Alarcon, ‘Cash Transfers and Poverty Reduction Employment Guarantee Schemes and Conditional Cash Transfers’ in A. Pankag (ed.), Right to Work and Rural India, Sage, New Delhi, 2012. Zepeda, E., L. Ndirangu, F. Leigh and J. Omollo, Kenya’s Youth Unemployment: the transition from child to adult in a rapidly changing country, 2013.

8 Policy, Institutional and Methodological lessons from six country studies Albert Berry Introduction The six case studies covered in this volume involve a wide range of country settings for social protection policies and for the interfaces of those policies with labour markets and hence labour outcomes. They also feature a wide range of types of social protection policies, of total resources directed towards them, and no doubt also of total impacts on welfare, poverty, employment and other important variables. It can be argued that the bulk of the reduction of income poverty in the developing world is due either to reasonably pro-poor economic growth or to social protection policies, with growth usually the main single factor at work, especially in lower-income countries. The challenge in designing good social protection policies is thus inherently tied to the nature of the growth process. A successful package of economic and social policies must produce (i)Â a good rate and pattern of economic growth to reduce poverty directly; (ii)Â a well-designed system of social protection to defend those still left in poverty despite the growth achieved; and (iii)Â internal consistency between these two broad categories of policy, such that neither cancels out the positive effects of the other. It is particularly crucial that social protection policy be as consistent as possible with the creation of good jobs, since it is mainly through job creation that growth contributes to poverty reduction.

288

Social protection, growth and employment

Policy makers and analysts should view social protection together with economic policy, to take full advantage of the potential synergies between the policies that can generate economic growth (macroeconomic policy, trade policy, industrial policy etc.) and the social protection policies whose main objective is to offer protection against some unacceptable outcome. The first set of policies can alleviate poverty, economic insecurity and attendant problems through growth; the second set responds to situations and people for whom growth does not provide an adequate outcome and/or economic security. While the above generalizations provide a useful starting point, differences across countries in level of development, structure of the economy, administrative capacity and other features mean that the design of pro-poor growth and effective social protection programmes is always quite country-specific. Although economic growth is almost always a key factor in poverty reduction, this is especially true for low-income, often mainly agricultural countries, which are generally also characterized by a low public spending capacity (due to low taxing capacity) and a low implementation capacity. For them, it is especially important that social protection policies be consistent with the growth process. In countries with high fiscal and implementation capacity, social protection can be a more potent tool against much poverty and vulnerability. Of particular interest to us here are the social protections that overlap with labour market processes and hence affect labour outcomes. Some of these raise the productive use of labour and hence GDP (as with India’s National Rural Employment Guarantee Programme), while others may directly diminish income earned through employment, such that the net benefit due to the social programme is less than the transfer involved. Another relevant distinction is between social protection whose income-creating effect occurs immediately — for example, employment-generating public works programmes—and that whose income-generating effect comes later, as with conditional cash transfers designed to keep the children of low-income families in school longer. Taking a broad perspective on how family welfare may be improved leaves us with the possibility that some of the best policies to ‘protect’ people against bad economic outcomes are ones that promote pro-poor growth. In that case, growth objectives and social protection objectives may fully overlap in the sense that the best policies for the first objective are also the best ones for the second. At the other end of the spectrum, the policies that generate fast growth may be exclusive, such that many people’s needs are not satisfied by that growth process and may even be damaged by it. Perhaps the two most important policy areas where growth and social protection are so highly correlated that they are virtually the same thing are small-farm policy, exemplified most clearly in the cases of Malawi and Tajikistan, and microfinance, exemplified in Peru. While small-farm support policy aims to raise productivity and incomes in the activity where most people, especially from lower income deciles, are found in the early stages of a country’s development process, microfinance has a parallel intent, namely to raise productivity and incomes of small operators mainly outside agriculture. In both cases success lowers poverty and economic insecurity directly through the increasingly successful activities of the producers themselves.

Policy, Institutional and Methodological Lessons

289

This important parallel notwithstanding, the two areas differ importantly in other ways. First, the small-farm support policy model has shown its success for literally hundreds of years, and has arguably been the most important mechanism of early economic growth (Johnston and Kilby, 1975) and a virtually indispensable feature in any early growth that was also equitable. In contrast, microfinance on a major scale is a recent innovation, dating back only a few decades, and hence is not time-tested as small-farm support policy is. A second difference is that microcredit (and credit for SMEs), as one element of support for smaller non-agricultural firms, may leave many of them with only modest potential to raise their productivity and incomes; although some microfinance systems have made serious attempts to deal with other needs of borrowers besides credit, it has not been standard practice to put as much emphasis on these non-credit elements of support. In the small-farm support model, in contrast, the central element is generation and diffusion of new technologies (Mellor and Johnston, 1961), primarily the responsibility of the public sector, with credit needs usually assumed to take care of themselves. A parallel process of technology provision has not evolved in the case of non-agricultural microenterprise. For one thing, microenterprise is a far more heterogeneous collection of activities than is agriculture, so the needed technological advances are also more varied. Further, it is the practice in industrial countries that outside agriculture private firms undertake the bulk of the applied research, and the role of the state is much less. Given the economies of scale1 that characterize many research processes, and the special advantage that accrues to innovations adopted by firms with market power, this has become the norm. Public-sector technological support for small farms has always made economic sense because there are very few economies of scale in agriculture, unless they are built into the research process itself (Lipton, 2009). Outside agriculture larger firms are more likely to have an advantage, not only because of the greater frequency of economies of scale but also because size often bestows market power and the resulting ability to charge prices above costs. These potential and often actual advantages of large firms, plus others, lend plausibility to the fear that microcredit may not be as central an instrument of poverty reduction outside agriculture as technological support for family farms is within it. If technological barriers keep the productivity of small firms below those of larger ones, then, in contrast to agriculture, their ability to compete may be present in too narrow a range of activities to allow the MSME sector to be a major motor of equitable growth. MSMEs will crowd into those sectors without important economies of scale (some types of retail commerce and of services, for example) and push down the relative prices of the goods and services involved in such a way that the benefits wind up going mainly to others. This interpretation contrasts the smaller technological range that existed in the earlier stages of development in the now industrial countries with the bigger range of today, and attributes the prevalence of urban informal workers, dualism and income inequality to it. A third area of clear interface between social protection and growth policy is education, where conditional cash transfer (CCT) programmes aim at keeping the children of lower-income families in secondary school longer (this now being the frontier level in most countries where such programmes are in place). For this social protection to fulfil expectations, the private and social payoffs to increased levels of education for the groups in question are important. It is now highly probable that the earlier very optimistic readings on the payoffs to more education (both

290

Social protection, growth and employment

private and social) were exaggerated—how much so remains a matter of debate.2 The mediumand longer-term benefits of these programmes may fall well short of hopes and expectations if either (i) overall returns to education are now on the low side; (ii) returns to secondary education are particularly low and the CCT programmes do not provide a bridge to higher education for many of the benefited students; (iii) the quality or appropriateness of content of the schooling provided to the beneficiaries is low such that the payoff to them is lower than the average for the level of education in question; or (iv) there is discrimination in the labour market against these graduates such that their education cannot pay off as well as it otherwise might. The current dissatisfaction, frustration and protests with respect both to education (cost, quality, access) and to employment options of youth and others (e.g. Kenya) show that there are no easy fixes in this area. It appears probable that for at least several decades into the future a high share of graduates of the educational systems will move into informal activities. This fact is not well recognized by many educational decision-makers, nor by many students and their parents, suggesting the need for a major rethink which, while ongoing, remains incipient in most countries. Another important area where growth and social protection are often pursued together is labour-intensive public works (as with India’s NREGA). Pensions and other transfers that get into the hands of people in low-incomes area with much underemployment can have positive multiplier effects (also analysed in the India case study). Health interventions are expected to raise current productivity when targeted to members of the labour force and future productivity when targeted to children. It is particularly important in Tajikistan and in Malawi, the two very low-income countries among the six cases studied here, that social protection policies be consistent with and foster the present and future motors of growth in the country. That the most important way to achieve social protection is with a good growth and employment strategy follows partly from the centrality of growth policy but also partly from the fact that most forms of service provision are difficult in such low-income, rural economies with limited administrative capacity. To do well in the medium and longer term, Tajikistan clearly needs both to turn small-scale agriculture into a bigger source of income than it now is and to find new economic directions to provide future motors of growth and employment creation. An analysis of employment and labour productivity is thus of central importance to understanding the opportunities and constraints faced by Tajikistan. In Malawi, agriculture not only is but will continue to be the key sector for some time.

An overview of country results The six countries analysed in this volume represent a wide range of per capita income levels and associated characteristics and differ in other important ways as well. Malawi is a low-income, highly agricultural country with a continuing very fast rate of population growth (2.8 percent over 1998–2008), life expectancy of about 57 years, adult literacy of 72 percent, a modest level of current educational attainment3 and a GDP per capita of US$761. Land is very scarce, with a mean farm size of 1.2 hectares (or 0.33 ha. per person).

Policy, Institutional and Methodological Lessons

291

The public sector and its capabilities are very constrained by the low per capita income, albeit somewhat relieved by a high ratio of foreign assistance to public spending. Although other sectors must gradually take on increasing importance, success in poverty reduction is likely for some time to depend mainly on an increasingly productive agricultural sector, whose benefits are widely distributed due to the relatively equitable access to land. Although other social protection policies should be and are present, they do not compare in importance to raising agricultural productivity. In this case, therefore, a good pro-poor growth policy is also by far the best social protection policy. Malawi’s still smallish urban sector comprises nearly all of the country’s formal employment, and the majority of this is in the public sector. Issues of urban informality and poverty, while important, do not yet compare in seriousness with rural poverty, nor should they be given the weight in policy that must be directed to the latter. The Malawi case study (Chapter 2) provides strong evidence of the poverty reduction impact of maize technology transfers, initially effected through the free distribution of small start-up packages and later through subsidies. The maize growth boom over 2005–2010 significantly reduced the need for food aid and safety-net measures to mitigate hunger. The welfare benefit was well distributed across all strata of poor people; the non-poor gained from increased crop sales and increased home production, while poor people increased subsistence production and also gained from higher wages for their casual labour, as the growth of production (in both the targeted crops and indirectly affected traditional crops) fully absorbed the rural labour supply throughout the main cropping season and created labour shortages leading to substantially higher piece work rates. Input transfers and subsidies along with other social protection measures also facilitated the empowerment of women, both directly and indirectly—through the opportunities afforded to expand production of crops in which their entitlement is secure. In fact, the output growth of these crops, which directly benefit the household (food) and themselves (income), exceeded that of maize and tobacco. The smallholder response to input transfers and, especially, the Agricultural Input Support Programme (AISP) confirmed the relative productivity of small farms (compared to the estate sector/larger smallholders) and their greater social efficiency in utilizing available labour resources. The various programmes have also collectively advanced smallholder knowledge of the value of new technologies; in the case of maize, this has resulted in more widespread adoption of modern seeds and displacement of low-yielding traditional varieties. Over this brief timeframe, poor and ultra-poor people have managed to accumulate productive assets—most notably livestock (goats, pigs and cattle)—and also to set aside savings. The increase in small livestock is especially important because women are culturally entitled to possess goats, pigs and poultry, so the growth trend implies that their livestock asset position has correspondingly improved. All of this suggests the potential for a very positive long-term impact if momentum is not lost. Among the practical lessons to emerge from this experience is that input transfers and subsidies were most effective when targeted broadly—i.e. to virtually all small farm households— partly because narrow targeting is difficult, partly because nearly all farmers are both relatively small and relatively poor, and partly because to the Malawian small farmers themselves, universal targeting is justifiable as neither farm size nor income vary greatly. The effects of ‘leakage’

292

Social protection, growth and employment

through imperfect targeting are alleviated by non-formal mechanisms of wealth redistribution (or reciprocal gestures) in this strongly traditional society. The benefits to smallholders have been widely distributed, both directly (through sharing assets/inputs) and indirectly (through provision of employment opportunities). Because of its high fiscal cost, the size and scope of the input subsidy programmes have been reduced to focus on maize alone. Subsidization on a large scale requires sustained donor financial support in order not to excessively shrink the rest of the agricultural budget. But available research indicates that only 20 percent of rural households currently have the ability to procure a seed and fertilizer package at the full market price, while about half could afford a 50 percent subsidized package. Reduction in fertilizer use could reverse the gains recently achieved. With an initially very high level of poverty and with the optimal paths to growth and to poverty reduction essentially the same, it is clear that Malawi was on the right track in strongly supporting yield-enhancing developments for small farmers. For agriculture to maintain its role as a driver of economic growth and poverty reduction, an appropriate basket of growth-cumsocial protection measures should include a universal inputs transfer to poor and ultra-poor people who have land and labour to increase household maize production, an inputs subsidy for the non-poor for investment in either maize or commodity crops, and both food and cash transfers to specific ultra-poor cohorts, especially the landless, households unable to work and those households without a supply of labour. The government endorsed this strategy in its 2009 Social Support Policy. But whether it can stay the course and convince donors to stay on board is a political question, subject to domestic political pressure (from poor people, who do not want to see their gains reversed, and non-poor who want to retain their position as beneficiaries) and donor equivocation over policy direction. The potential role of other forms of social protection, including public works programmes and direct transfers in the reduction of poverty is unclear; although public works programmes are a logical complement to rising crop production because of complementary seasonality and the need for better infrastructure. Their impact has, however, been geographically limited, and their main beneficiaries have been families who have in any case gained (both directly and indirectly) from the input transfers and subsidies. Given the narrower objectives of some of these alternative social protection measures, their main impact has probably been to smooth consumption, although their role in empowering marginalized smallholders (the ultra-poor) and women must be noted. In short, the experience of Malawi confirms the great potential of a small-farmer based pro-poor growth-cum-social protection policy; the surge of yield, output and income increases based on subsidized inputs could be the first, albeit interrupted, step in a virtuous circle spiralling upward if the policy can be reinvigorated and a local research and development apparatus built to provide the next steps in pushing yields upward to the point where farmers would be both able and willing to fund their inputs without subsidy. Without those next steps there is the serious possibility that the gains achieved will prove to be one-off in nature or worse still, that yields will fall back and not quickly regain the peaks they did reach, while population growth

Policy, Institutional and Methodological Lessons

293

continuously shrinks the already meagre amount of available land per person. Unfortunately the latest chapter of the Malawi story is not a happy one, as successful policies have not been maintained. This wrong turn probably precludes for the time being the possibility of an acceleration into sustained growth, and may well cancel out all of the benefits of the recent growth spurt, except the knowledge that such an experience can happen under the right circumstances. Peru, a middle-income developing country, has experienced good growth over the last two decades, following a lengthy period of erratic and generally unsatisfactory performance. Urbanization is well along; life expectancy relatively high and rising. The country suffers from the dualistic economic structure of most Latin American and many other countries. Income inequality and asset concentration have historically been high and ethnic income differences marked. A long history of dependence on mineral exports is one aspect of the dualism and presumably a contributing factor to the high levels of inequality and to the informality of urban employment. Recent analysis and debate around the trends in inequality feature an optimistic view that inequality has been falling, along with some scepticism on this point. The prominent social protection issues in Peru are characteristic of middle-income, highly urbanized countries, where access to the more ‘desired’ jobs is difficult for many and where informality is high and poverty prevalent. The response in terms of social protection programmes and policies is also similar to those of other developing countries, especially in Latin America. In several respects, however, the country’s economic and/or policy experience has been unusual or extreme. As a mainly mineral exporter, it has suffered severe economic swings related to those of the international markets for its products and to its own major policy swings related to its political instability and the widely differing ideologies of successive governments. At various points from the 1950s to 1990 policy was relatively interventionist in areas such as trade and the role of public enterprise. In the 1980s its labour legislation was more restrictive than the average for Latin America, which in turn was more so than for the developing world as a whole. The economic crisis suffered in the late 1980s was very severe, with GDP per capita shrinking by 28 percent over 1987–1990. Policy then shifted sharply in a market-friendly direction under president Fujimori. The crisis and the dramatic increase in poverty which it produced focused attention on the need for much more active social protection policies as the 1980s crisis of heavily indebted countries that spawned the so-called ‘lost decade’-had in many other countries of Latin America and the developing world. Peru’s situation was one of the more extreme, both in terms of the economic shock involved and of the abruptness of the country’s shift to a markedly lower level of interventionism, including a major reform of labour regulations in the early 1990s. Another special feature has been the extensive development of microfinance, in part a natural response to a very large informal sector but in part due to policy support. While the precise impacts of the 1990s labour reform remain uncertain, it is clear that it did not lead to a large and quick reduction in the size of the informal sector through the predicted flexibilization of formal-sector labour contracting. Peru not only maintained one of the highest levels of informality in the region but also experienced a consistent growth in the share

294

Social protection, growth and employment

of informal employment both during the late 1980s crisis and during much of the subsequent growth recovery of the 1990s. The expectation of a quick and significant increase in formalsector employment as a consequence of the reform seems in any case to have been unrealistic in light of the relatively small number of firms on the margins of formality and capable of being formalized by such legislative changes.4 While the labour reform did not produce the substantial benefits hoped for by its proponents, it remains unclear whether it may have been responsible for more modest effects in one or the other direction. Informality rose until about 2000 but has since fallen somewhat and appears more or less in line with that of other Latin American countries, when per capita income is taken into account (ECLAC, 2004). It is thus not impossible that the reforms have given a small and delayed impulse to formality. The observed fall in labour income inequality over the past decade could also owe something to it. A best guess though is that the reforms had no major impact on either of these trends. As for microfinance, the main positive conclusions of the general literature show up in Peru as well, including an average positive impact on microenterprise profits (controlling for selectivity into the lending programme), on enterprise revenue, enterprise fixed assets, business premise ownership, business licensing and household income, as well as a capacity to reach many of the poor people. Its relative prevalence and the favourable conditions for its success make it plausible to expect that the benefits of microfinance would be large enough to be detected at the aggregate level on poverty, economic security, inequality and growth.5 Panel data for 2004–2008 (Rodriguez and Minora, 2010: 153) reveal a striking annual average growth in value added per worker of 4.1 percent. Some of this rapid labour productivity growth may reflect basic productivity enhancement. A next step, in understanding both the informal/microenterprise sector itself and the role of microcredit, will be to probe these possible links more deeply. Several of Peru’s social programmes per se stand out for the general lessons they suggest. Juntos, the transfer programme conditional on children attending school and the family taking advantage of health services, is generally considered successful on the grounds that it increases school attendance, monetary measures of income and consumption associated with a moderate poverty-reducing effect, and the utilization of health services by both children and mothers, and improves the nutritional intake of households. No impacts on final outcome indicators such as malnutrition or anaemia have been identified, which might simply reflect a time lag or, more serious, a need to complement it with a more adequate supply of quality health services. The Glass of Milk programme has the widest coverage among Peru’s food-subsidy programmes directed at women and children pre-identified as vulnerable to food insecurity. It is run through civil society-based organizations and operated by mothers. The local operation has been criticized for inefficiency and corruption, including the persistent opposition of beneficiaries and their leaders to graduate/exit and thereby pave the way for new participants; protests from this group have made the government keep changes in management and supervision to a minimum. The programme’s continuous operation has relied on close ties between the mothers in charge and the government authorities, which helps to explain its straying from its original goal of serving only women and children suffering current food insecurity. An important positive side effect has been the high participation of poor women, empowering them in economic as well as

Policy, Institutional and Methodological Lessons

295

political terms as leaders emerge from among them; it is cause for some concern, but perhaps inevitable, that most of the leaders are not from the poorest quintiles, which tend to avoid such responsibilities because of lack of time or interest. Although targeting has not been perfect, it has had good outreach among the poorest populations (quintiles 1 and 2) and those with low nutritional status, a feat attributable to the targeting of poor districts and the fact that poor people received larger in-kind transfers due to intra-district targeting. However, no clear impact on nutritional status has been detected; according to some analysts, this is because the transfers of milk and milk substitutes are infra-marginal for about half of the recipient households. Although this does not imply that the programme has a low or zero payoff, it does show that to achieve a major nutritional impact requires even more precise targeting than simply getting milk to families who fall under a specified monetary poverty line. The Integrated Health System’s (Sistema Integral de Salud—SIS) aim of providing fully or partially subsidized health insurance to poor and extremely poor people has been at least partially achieved over its decade of operation. The system relies on the existing public hospitals and the public health centre system to provide health services, in return for which these institutions are paid by the SIS system. Coverage has been continuously broadened to include more complicated health treatments and services. Information campaigns and the simplicity of enrolment have multiplied the number of beneficiaries quickly, but at a rate that neither the budget nor the supply of quality services has been able to match. By 2010, the programme reported over 12 million insured people, or more than a third of the country’s population, with two thirds coming from the poorest 40 percent. The great majority of beneficiaries (95 percent) sought health services in health centres and posts. Many complaints have come from the client population because of limited and slow medical attention, and many more from the public hospital administrators and professionals because of payment delays and underpayments from the SIS programme. The combination of the enormous increase in the number of users, no additional budget for investment in physical infrastructure and only minimum increases in that for medical professionals and medicine, along with serious underpayment of their contributions by beneficiaries, has in effect created a huge excess demand for the system’s services and an enormous fiscal problem. The immediate consequences have included a de facto lack of coverage for an important proportion of the targeted population—in particular, the poorest families, many of whom are located in small towns and other places not yet included in the programme—and a decline in the quality of health services. These are a natural, probably unavoidable result of a sudden surge in demand, which under the best of conditions would be hard to cope with, but which is aggravated by underlying inefficiencies—inadequate monitoring of this demand-side subsidized programme, deficient information systems and the weak capabilities of managers. Given the inevitability of a gestation period during which such a system ‘learns’ how to operate effectively, it is important to learn from experiences such as this one. Clearly, overloading the scarce health professionals can be counterproductive in the medium term. There are likely to be trade-offs between targeting the poorest people first, keeping costs down and keeping professionals satisfied. Peru, with its rapid increase in public revenues from mineral-based growth, has a better chance to meet these challenges than many other countries.

296

Social protection, growth and employment

Provias Nacional is a large programme of transport infrastructure projects that works with private-sector firms and aims to generate a considerable amount of local employment. The economic logic of building or enhancing rural roads that help to connect low-income farmers to the market has long been recognized, especially when the cost is modest and much of the labour can be provided locally. Beyond its road-related benefits, Provias also appears to have had important effects on local employment and possibly on small local economies. One evaluation pointed to significant improvements in school enrolment, visits to health centres, land use for farming, and male agricultural salaries. Rural entrepreneurship was strengthened, and after initial male resistance was overcome, women came to account for a quarter of the staff in the microenterprises. A valuable process of learning through experience has contributed to Proviasâ&#x20AC;&#x2122;s success. Institutional improvements, increased local participation and better local governance were all encouraged by the project and proved beneficial. For example, the fact that a measured impact on poverty was delayed by several years after roads were rehabilitated provided part of the incentive for the creation of a Local Development Window managed by NGOs to identify and prioritize local productive initiatives. Involvement of women in the design of projects was critical to meeting their needs, such as construction of footpaths, preferred by women because they are the safest way to take animals to pasture and to collect firewood and water. According to the World Bank (2010: 5), this programme has become a benchmark for best practice in other rural transport operations in Latin America. Provias is of interest also because of several methodological complexities surrounding the confirmation of its effects. One involves measuring the benefits to women from the nonmotorized roads, given that many or most of these did not take a monetary form.6 Another is that the time trajectory of benefits may be very important. Thus some studies of Peruâ&#x20AC;&#x2122;s investments in rural roads have found no general tendency to resulting income gains in the short term, although favoured beneficiaries near motorized roads did clearly benefit. A final desirable form of analysis around such a project, where the initial micro-level evidence is encouraging, is a look at how big such an intervention could become. Such an assessment entails considering how much more rural infrastructure investment could be undertaken with comparably positive effects. Tajikistan, a low-income former member of the Soviet bloc in transition towards a more market-based economy, has been disadvantaged by the severe civil war that followed the breakup of the Soviet Union and by unfortunate policy-making, in particular the failure to expeditiously distribute land into small units and to support the new small farmers. On the other hand, its citizens have had access to Russia as a destination of emigration, and the resulting remittances have formed a very large component of the incomes of many Tajikistani families. Since, however, remittances can be an unstable income source and are not available to all families, the short- to medium-term challenge involves developing a system that complements and eases the welfare problems of such a remittance-focused system, while also addressing the needs of the families for whom remittances are not a significant resource. As in Malawi, the other very lowincome country in our sample, it is particularly important that social protection be consistent with and foster the present and future motors of broad-based growth in the country.

Policy, Institutional and Methodological Lessons

297

The evolution away from a Soviet-style system with its large enterprises and centralized control presents several other major challenges. First, the typically low labour intensity of larger units in market systems means that privatization shrinks employment; to make things worse, the former system does not provide a natural base from which more employment-creating small or medium firms can quickly emerge to fill the gap. Second, in the hands of new owners/ managers inclined to take advantage of the power that a large enterprise can wield, the system becomes monopoly-ridden, further discouraging the creation of small firms that could otherwise be expected to enter these markets. Instead of state monopolies intended to be managed in the public interest, private ones emerge that serve the narrow interests of those who control them. The transfer of public to private wealth during the transition of former Soviet bloc members has underlain much of the sometimes huge increases in inequality that resulted. In terms of its economic structure and employment, income and economic security challenges, the transition has thus brought Tajikistan closer to the structure of a market-oriented economy whose comparative advantage is in natural resource products and whose institutions of economic and social policy-making are weak and controlled by a narrow elite. Thus far, the failures of Tajikistanâ&#x20AC;&#x2122;s economy itself have been substantially alleviated by the large flow of remittances from emigrants working in Russia. It is unclear how long this will remain the case if the country continues to fail to generate adequate jobs domestically. The most important component of a strategy to achieve healthy pro-poor growth and continuing poverty reduction is to advance the family-farm sector by further allocating land into small units, reforming the marketing systems for agricultural inputs and outputs and developing a support system that can provide the infrastucture and new varieties and breeds needed to raise productivity. Successful development of small-scale family agriculture is especially critical to improving food security and reducing poverty. Despite its limitations, land reform has played an important role in the limited agricultural recovery achieved so far. The performance of household plots has been markedly superior to that of the other tenurial forms. But the failure to move it along expeditiously has contributed to the vicious cycle hindering the countryâ&#x20AC;&#x2122;s efforts to pull its population out of poverty. Cotton is a large component of agriculture. Independent Tajikistan inherited from the Soviet era the remains of a once thriving sector, organized into large units. Here, as elsewhere in agriculture, the challenges are to move as rapidly as is practically possible towards greater ownership and control of land by small farmers (and their communal associations), the development of a decent support system and the dismantling of the monopolistic control of land and related assets by an in-group which has been blocking these processes. There have been halting steps in this direction. With Tajikistanâ&#x20AC;&#x2122;s agricultural sector under the Soviet system not geared towards food, independence led to a severe supply crisis, so in 1992 the government passed its first of several land reform packages aimed at alleviating that crisis. While pointing in the right direction, these reforms and the related process of land transfer have been far too slow, delayed among other things by the opposition of high-level bureaucrats from the former state-controlled system and by land allocation processes that favour the well-connected. The monopolistic nature of the cotton market and its adverse effects on investment and growth are widely acknowledged, but reform is politically difficult, and the cotton-growing areas, suffering

298

Social protection, growth and employment

by far the highest level of extreme poverty, remain the source of a majority of migrant labourers. Policy recommendations need to be structured around the principle of achieving an adequate level of local food security, by allowing crop diversification to modify the damaging aspects of cotton monoculture. As a relatively labour-absorbing crop, cotton has the potential to contribute importantly to pro-poor growth. To achieve that pro-poor growth and adequate employment creation, Tajikistan must be successful on a number of labour-intensive fronts, including cotton, other agricultural products, labour-intensive services and, possibly, some branches of manufacturing. After agriculture, support for non-agricultural micro, small and medium-sized enterprises (MSMEs) is the other main sector-specific policy needed. Growth of smaller-scale MSMEs has been considerable, but this is probably due as much to the lack of alternative sources of employment as to any positive attraction. Despite some growth in the SME sector and the expectation that it will gradually become more formalized, the informal sector is likely to remain pivotal for employment generation for several decades; prohibitive measures against it would thus be counterproductive, possibly seriously so. Tajikistan faces many of the challenges to the success of this sector that emerge in market economies, including a financial system that is not very effective at supporting SMEs and a degree of negative reaction on the part of bureaucrats and others to the informal sector. Non-agricultural rural activity is severely constrained by inadequate access to energy and electricity. Probably the most important policy step related to this sector is to avoid penalizing small firms, including microenterprise, through either unnecessary regulations and requirements or more punitive measures (IMF, 2010). Unfortunately some measures of this sort have been taken, such as the closing of small markets in cities where small producers come to sell their agricultural products. Attempts to discourage informal activity through high taxes and other impositions are likely to raise unemployment and push more people to emigrate. One necessary step is to provide a simplified regulatory system, including easier procedures for registering small businesses, tax cuts and possibly exemption from taxes during a period after business startup. As well as making the bureacracy less of an impediment than it has sometimes been, success will also call for developing a positively supportive set of programmes and policies. Measures to promote small businesses can contribute both to legalizing part of the informal sector and improving its accountability in terms of minimum social guarantees. On this proactive side, one useful step is to support the involvement of small and medium-sized businesses in the process of government procurement, which can be facilitated by reducing the batch size of public contracts to ensure that they are not squeezed out. Development of the countryâ&#x20AC;&#x2122;s hydroelectric potential is a high priority both as an input to industry and as key to the provision of electricity and water to the population. Since they create few jobs, such capital-intensive activities must, however, be balanced by labour-intensive ones (such as family agriculture and MSMEs); otherwise any growth that results will be neither employment-creating nor pro-poor. The health and educational systems and the infrastructure behind them are also significant challenges, given the deterioration they have suffered.

Policy, Institutional and Methodological Lessons

299

Finally, a new social protection system that complements pro-poor economic policy will need attention for the foreseeable future. Dissolution of the Soviet Union and the severe economic crash that accompanied the civil war affected both Tajikistanâ&#x20AC;&#x2122;s labour market and its social protection system in very negative ways. The previous arrangement, with the state responsible for employment, job security, pensions etc., provided a good defence against extreme poverty and economic insecurity. The defence mechanisms against poverty and economic insecurity that grow up in a market system (see Chapter 1) are not initially present during a transition from central planning to markets, and their evolution tends to be slow, since their effective functioning calls for expertise and institutions specific to the different needs of the new system. The drastic deterioration of the previous social protection system has seen the consolidated budget for social assistance (excluding social pensions) fall to a miniscule 0.2 percent of official GDP in 2009 (World Bank, 2010: 1). Because this assistance is not well targeted, it exerts almost no downward influence on poverty rates or on income inequality. The other main form of social assistance, managed by the state agency for social insurance and pensions, takes the form of pensions for poor elderly people who have not contributed to a pension fund (ibid.: 8). This amounts to 0.3 percent of official GDP. In broad terms what is needed is a major increase in spending plus a good targeting system to replace the rather accidental processes currently in place. The new system will need to be carefully designed to avoid onerous regulations that would impede the successful evolution, including eventual formalization, of smaller private firms. Among the priorities are food security (beyond that implicit in a strong family farm system), improvements in primary health care, and programmes to assist the poorest and most vulnerable people. Mexico has the highest per capita GDP of the six countries studied, but the second slowest growth over the last 30 years (exceeding only Tajikistan, which suffered a civil war) at just 2.2 percent per year. It also has the most open economy, measured by export/GDP and import/ GDP ratios, related in part to its proximity and special trading access to the US economy. The growth in the labour force reflects both the very fast vegetative population growth of the mid20th century, gradually slowing to about 1.5 percent now, and the massive net emigration to the USA over recent decades. The emigration, mainly of manual workers, has lessened the supplyside pressure on that component of the Mexican labour market, while having also a variety of other effects. As befits the country with the highest income in the sample, Mexico has the widest assortment of social protection policies. Together with countries that broadly share its level of development and labour market features, especially its high level of informality, it provides a laboratory within which several important challenges are being played out. These include (i)Â the impact on formal-sector employment growth of labour market reforms aimed at making that market more flexible; (ii) the impact on formality and on total employment of broadening coverage of social programmes that provide some of the same elements of social protection as do formal-sector social security arrangements; (iii) the direct social benefits of widening the coverage of the social protection programmes mentioned; and (iv) the success of programmes designed to raise the productivity of the many enterprises that have remained and will remain in the informal sector.

300

Social protection, growth and employment

The Mexican experience is especially relevant to the last three of these challenges. A major debate has developed around the possible inefficiency of the combination of traditional formalfirm-based social security provision and the population-wide coverage of some recent types of support, and around the extent to which these latter impinge negatively on labour market performance. Levy (2008) has argued that the widening coverage of social protection programmes encourages informality; studies such as Knaul et al. (2012) contest that view. It does appear that the level of urban informality in Mexico is somewhat above that of other countries of similar per capita income.7 Data presented in Chapter 5 on worker movement among firms indicate that the flows between units of smaller and larger size are limited, which may suggest a modest impact of social protection arrangements on the distribution of labour between formal and informal activities. In any case, this important issue clearly requires more research to sort out the effects of existing programmes, and more effort directed to improving design as well as experimentation to find better versions of the existing social programmes. The Mexican case study provides a clear picture of the relationships between firm size, sector, type of work (permanent vs. transitory), on the one hand, and worker benefits and the reach of many government programmes on the other. It shows that not all services commonly thought of as being provided by formal-sector firms are in fact provided by all of them. More generally, it is evident that only a minority of Mexican workers could be described as having adequate working conditions, including a modicum of security, and that a large share of new jobs over recent years have been created in microenterprise. Recent trends lead Martinez (2012) to speculate that the informal sector is likely to remain large in Mexico for at least the next 50 years. Thus the matter of how social protection should be widened is of great importance, since it is clear that gradual growth of the formal sector with its associated social protection services will be too slow. If the informal (along with small and medium-sized) enterprises are likely to remain important employers into the medium and long term, how well are Mexican support policies supporting their productivity and growth? The case study clarifies at least one aspect of the picture: which services are reaching whom. The Small and Medium-Sized Enterprises Fund (Fondo PyME) is directed at five business segments: entrepreneurs, microenterprises, SMEs, ‘gazelle’ companies (those with potential for sustained high growth) and driver or magnet companies (whose presence can attract other firms to start up nearby). These segments receive support for finance, trading, training and consulting services (Fund Evaluation, 2007–2008). ENAMIN 2008 data show that just 4.7 percent of units aware of the service used it. The National Programme for Microenterprise Financing (PRONAFIN) places special emphasis on those with higher poverty levels and promotes productive initiatives of individuals and social groups in poverty conditions in every region of the country by encouraging and promoting a microfinancing system. Evaluations show that this programme helps to increase income in those units that received credit, although job creation was small or negligible, and coverage (at about 0.3 percent of small enterprises) was very small (PRONAFIN, 2012). Such findings suggest that the existing support system is not of sufficient scope to provide a strong impulse to productivity and income growth in the small enterprise sector of Mexico’s economy.

Policy, Institutional and Methodological Lessons

301

Mexico’s social protection policies range from some that appear to be successful in achieving their goals to others that do not or about which judgement must be reserved at this time. Like nearly all countries of Latin America, Mexico has historically suffered a high level of income inequality, a level that rose significantly over the economic adjustment period but has, by some estimates, begun to fall during the last decade (Gasparini et al., 2009). The long-term pattern of (income) poverty incidence saw a marked fall during the fast growth period up to 1980, followed by fluctuations since then. According to Lustig et al. (2011), income redistribution through the budget (i.e. through the structure of taxation and of public spending) lowered Mexico’s Gini coefficient significantly in 2009, from 0.502 to 0.433—considerably less than in Argentina (13 points) but much more than in Peru (2.5 points); they conclude that Mexico and Peru get the most redistribution per dollar spent that way. Though rough, this estimate leaves open the possibility that some programmes are significantly redistributive. India’s National Rural Employment Guarantee Programme (NREGA), launched in 2006, is the world’s largest social policy programme. Its enormous scope and strong economic logic make it an important test case. Its implementation has coincided with a period of very rapid growth from the 1980s, a sharp increase in inequality beginning at some point in the 1990s, a very small and slowly growing formal sector (Mazumdar and Sharkar, 2008), slower growth in agricultural wages since the early 1990s than before, and major implementation challenges presented by inefficiency and corruption. Given the continuing high incidence and depth of poverty in rural India, the success of this programme will matter to millions of people. Measuring the impacts of a programme such as the NREGA involves at least four major methodological challenges. One is to measure the coverage and direct benefits to the target families in a context known for high levels of corruption and misdirection of funds. A second is to evaluate the indirect impacts and hence the overall (general equilibrium) effects after allowing for any substitution phenomena—for example, less labour supplied to other possible uses due to the presence of the programme or positive multiplier effects, say from local spending of the income earned. The third is to put a reasonably accurate value on the structures built, reinforced or maintained by the workers in the programme. The economic value of these outputs, including elements of rural infrastructure, is of particular interest given the general recognition that lack of such infrastructure has been a serious barrier to growth of rural incomes in India over recent decades. Finally, it is important to be able to detect changes over time in how the programme works, either improvements as personal and institutional learning takes place or deterioration and corruption as people learn how to better exploit the programme for their own ends rather than those of the programme or inefficiency grows with time. The recently broached possibility of greatly reducing corruption within the programme by transferring funds directly from the sources to the recipient families by making sure that all have a bank account (The Economist, 2012) will be a major test of an approach which may also be applicable to various other social protection programmes.8 Earlier analyses of the NREGA report that, despite the many impediments to full success, positive direct effects on poor people in many regions of the country are clearly visible. Accepting this starting point, the present study uses a general equilibrium model to estimate NREGA’s

302

Social protection, growth and employment

indirect effects. The conclusions are generally quite positive. Most importantly, the programme generates a ‘virtuous’ redistributive effect, since the benefits accruing to the beneficiaries are not limited to the wages directly paid by the jobs created.9 Positive spillover effects lead to additional jobs likely to be taken by poor people in both rural and urban areas, thus reinforcing the benefits to poor people in rural areas and extending them to those in urban areas. Although the estimated increase in labour income from these additional jobs is small, this is due in good part to the programme’s still comparatively modest size. Second, the cost of the programme is manageable. Government expenditures are funded by an increase in income taxes paid by wealthy people in urban areas, which end up imposing at most small reductions in consumption and welfare on them. As a result, the programme’s immediate poverty reduction effect is achieved at no net cost to the economy more broadly or to GDP in particular; rather the increase in resource utilization, directly and through local multiplier effects, means that it can contribute positively to GDP. The study does identify one potential source of concern to poor people: as the economy expands and their lot improves, the price of the commodities they consume is likely to increase. Although it is small enough not to be a matter of much concern in the aggregate, since the programme’s benefits much more than compensate for it, for poor rural residents who do not benefit from the programme’s jobs, even this small rise in the cost of living could be important, suggesting the need to complement the programme with others aiming to provide complementary support to poor non-beneficiary households. The programme’s positive growth and distributive effects are not threatened if it fails to reach all poor people in rural areas; rather, they are likely to be roughly proportional to its budget, at least within the range of changes analysed in the case study. This suggests that, although it represents a significant administrative challenge, the programme’s rapid expansion does not create a risk of negative growth or distributive effects. Needless to say, the benefits would be greater if operational efficiency were increased. If an increase in land productivity or in efficiency of local commerce is assumed to result from the programme, the short-term employment creation and output benefits are accompanied by longer-term growth benefits and any indirect redistribution that results. Since the direct benefits of higher land productivity do not generally accrue to poor people, given the concentration of land ownership, the direct distributional benefits from this higher productivity may be negative, although poverty effects may be positive and the distributional effects through holding down food prices will be positive. Such negative effects as may arise do not constitute an argument against the programme, but they do suggest the need to reinforce its provisions aiming to improve productivity on the lands of poor people. Some of the programme’s employment creation and land productivity effects complement each other well, as in the case of offsetting effects on the price of food, rendering it a potentially powerful tool to decrease poverty in the immediate and long term. Given its potentially enormous size, a programme such as the NREGA calls for the sort of general equilibrium evaluation undertaken here, and the exercise confirms the usefulness of the methodology—allowing quantitative estimation of indirect effects whose presence may be deduced by other means but whose economic importance cannot be assessed in any other way.

Policy, Institutional and Methodological Lessons

303

Large social protection programmes in other countries would benefit from this sort of evaluation. Only when indirect effects are taken into account can a country accurately assess the effectiveness of its social protection strategy. Designing and implementing a successful strategy for employment, wages, job stability and working conditions (the labour outcome variables) is a huge challenge in many developing countries. A dramatic example of how difficult it can be to satisfy the needs of a rapidly growing and youthful labour force is provided by the experience of Kenya. Chapter 7 describes in detail the structure of two key labour variables—open unemployment and wages—in the rural and urban areas of the country. There is no doubt that open unemployment is a major problem in Kenya, particularly among young people. As of 2005/06 it averaged 10.5 percent in this still mainly rural country, and youth unemployment exceeded 30 percent for those around 20 years of age. It is significantly higher among young females than their male counterparts—by more than 10 percentage points for those aged 15 to 25 years; at its peak for women a little under 35, unemployment approaches 50 percent. Females evidently have fewer job opportunities. At the other end of the spectrum of job outcomes, formal-sector employment is limited, at about 14 percent of the total in 2005/06. Though few, these jobs are the preferred object of search for many young people and others. Meanwhile, wage earnings vary widely by level of education. Tertiary education has increasingly become the norm for workers in the formal sector. The Kenyan labour market shares many features with the majority of developing countries, where youth unemployment is substantially higher than adult unemployment, urban unemployment higher than rural, and female unemployment higher than male; earnings differentials by level of education are relatively large, especially between those with tertiary education and others; and average earnings are substantially higher in the formal than in the informal sector and in urban areas than rural ones. It has long been hypothesized (Harris and Todaro, 1970) that large earnings differentials in favour of formal employment explain a substantial part of urban unemployment, as young people find it more sensible to search for such jobs than to look for informal activities or to remain in rural areas. It has also been argued as far back as the early 1970s and in the context of Kenya (ILO, 1972) that informal activities can make a major contribution to any economy and may be the only option for the majority of young people for at least some decades into the future. It has further been argued that the expansion of education in the absence of a growing formal sector will lead to a major mismatch between the jobs young people desire and those available to them, with a high cost in frustration and possibly also in economic efficiency. This view may even suggest that the expansion of educational attainment should be slowed if the growth of remunerative jobs cannot be expanded faster. Certainly it implies a need for better linking between education/training and the sort of jobs that are available. Although it shares many features with other countries, Kenya is an extreme case, partly because of the exceedingly fast population growth over recent decades creating an unusually large youth cohort relative to overall population, with corresponding pressure on the economy to create many jobs quickly, despite the modest rate of economic growth. Job prospects remain

304

Social protection, growth and employment

reasonably good for those with tertiary education, with their significantly higher earnings presumably reflecting that employers value their skills. But the swelling ranks of young people with secondary-level education are faring much less well, and their immediate future will probably be the test of how well the country is able to deal with the employment challenge. Kenya has experimented with a broad set of employment policies, going back to the 1970–74 Development Plan (Republic of Kenya, 1969), which featured employment and unemployment as important policy issues, through to the Sector Plan for Labour, Youth and Human Resource Development Sector (2008–2012), which looks into unemployment as a structural issue, explicitly focusing on young people. The last 10 years have seen the implementation of a number of youth employment policies with an emphasis on public works, entrepreneurial development and skills enhancement, including KaziKwaVijana (KKV), which aims to create 200,000 to 300,000 jobs by implementing manual-based small projects. One component of the KKV aims at increasing the employability of young people by providing skills through internships and training in the private sector. Several major practical issues arise in the context of Kenya’s employment policies. One relates to the payoff to youth employment projects, typified by KKV. Such programmes are no doubt part of a good policy package, especially to the extent that they are effectively managed and also generate useful products such as public works and develop useful skills for later employment. While valuable, they cannot be expected by themselves to make a big dent in the employment challenge, partly because it is very hard for them to cover a large proportion of the young people looking for work and partly because they are by nature temporary. After a young person leaves such a programme, he/she still needs to find paid employment; otherwise, the programme has simply delayed the day of reckoning. To increase post-programme success, such efforts need to create useful skills, as these ones do. What are the useful skills in the near future of Kenya’s economic development? It is clear that a large share of net new jobs will be in the informal sector (including small-scale farming and non-agricultural activities). This suggests that the employment challenge will not be adequately met unless this sector raises its productivity while its employment is (necessarily) expanding. The largest single component of this sector is agricultural self-employment on small farms, so the success of productivity-raising efforts in this sector may be the main single determinant of overall employment success. Non-agricultural informal activities will also play a key role. How should education evolve in Kenya? In the worst of cases, the number of graduates with a strong preference for formal jobs and with an education/training background suiting them only for such jobs will rise quickly, while the number of such jobs expands only slowly. The mismatch between the supply of and demand for such people will then continue to grow, with corresponding frustration on the part of the young people concerned. Alternatively, the educational system may orient itself more directly to developing the skills that will be needed, many of them in the informal sector, as well as attempting to create a more realistic set of job expecta-

Policy, Institutional and Methodological Lessons

305

tions for graduating students. This approach will entail a greater degree of contact between the educational system and the economic world than now exists in most countries. Arguably the main implication of Kenya’s situation is the urgent need for a coherent and sophisticated policy on labour demand (job creation, including self-employment), involving both the formal and the informal sectors and with necessary emphasis on how to raise demand in the sectors that currently employ the bulk of the labour force. Although the allocative function of the labour market might be improved, leading to some increase in jobs, and the supply side (skills development etc.) can no doubt be improved, it is likely that the evolution of labour demand will mainly determine the labour market outcomes (employment, wages, work conditions etc.) over the foreseeable future in Kenya. Of the six countries analysed here, Kenya is arguably the most challenging both to analysts and to policy makers. The answers are not clear-cut, and more in-depth analysis and more effective use of such analysis in policy are needed.

Conclusion A number of the conclusions of this study are specific to the country analysed; their generality can only be the subject of hypothesis until comparable questions are asked in a wider range of countries. Such tentative conclusions/hypotheses are, however, an important building block in this area of analysis. A number of lessons have emerged from the case studies; they are divided here according to whether they relate to a policy area itself, to institutions or to methodologies that help us to better understand the issues. Among the policy conclusions that emerge strongly from the case studies are: i) Small-scale agriculture has a key role to play, best exemplified by Malawi, Tajikistan and Kenya, and to some extent India, as a motor of growth, a generator of employment and incomes for many poor people and the major source of food security for many families. In Tajikistan the thus far very partial and hence relatively ineffective transition from large farms to smaller ones is one reason for the continuing poverty and the employment crisis in that country. Malawi’s story includes the impressive effects of the fertilizer subsidy. In various cases there is the issue of how successful research and development has been/could be in fostering dynamism in this sector. In some, most obviously Tajikistan, property rights are a problem. ii) Food insecurity is a major problem for many poor people which needs to be taken into account in the planning and implementation of social protection programmes, especially in low-income countries that depend heavily on the agricultural sector, such as Malawi and Tajikistan. Good policy in that area begins with a strong smallfarm system and should then be complemented by a range of instruments from India’s rural employment programme to Peru’s Glass of Milk.

306

Social protection, growth and employment iii) There has been a dramatic failure of labour demand to match the supply in some countries, with resulting high levels of unemployment and underemployment, especially among young people as in Kenya but possibly appearing in Malawi in the near future, given its rapid population growth. In middle-level developing countries such as Peru and Mexico this inadequacy is reflected more in high levels of urban informality than in open unemployment. The need for and absence of holistic employment strategies is apparent in all six countries. It is due in part to the necessary complexity of such strategies, in part to the mismatch between current institutional structure and somewhat more suited to the pursuit of this goal, and in part to a lack of due recognition of how important a strong demand for labour is to achieve propoor growth. iv) The possibly complex interactions between labour policy, the structure of social protection systems and informality are not currently well understood but need to be. Mexico has been a main centre for the debate as to whether certain forms of universal social protection contribute to informality by lowering incentives to formality, with attendant costs to the economy and the society. The evidence on this remains ambiguous. v) There is an immense challenge to get things right in the area of education and training. While no one doubts that investment in knowledge and skills contributes to economic growth and that interventions such as traditional cash transfers can not only raise the level of human capital but also contribute to greater equality of both opportunities and outcomes, some recent economic analysis has also suggested that the average economic payoff to raising educational levels may be considerably less than has traditionally been believed (Rosenzweig, 1999; Pritchett, 2001), that the total cost of investment in education is now high in many countries, and that in some cases (Kenya stands out) there is an actual or looming mismatch between the expectations of young people graduating from school and the capacity of the economy to create the desired jobs. Tajikistan presents a different case; there is a concern there that since families and their children do not see a clear payoff to more education, they drop out of school. This results from the combination of the relatively low educational levels required of emigrants to Russia to get jobs there and the lack of clarity as to what sorts of jobs will be available within the country over the coming decades. Both education/training policy in general and the CCTs that contribute to it need continual rethinking to address these challenges. The Peru case study reports promising results, although there remains the underlying question of how large will be the payoff to the increases in education that result from the programmes. Mexico has been a pioneer in such programmes, and India is experimenting in the area. vi) The requirements of health policy stand in striking contrast in many ways to those of education. Here there is no question that needs are high and greatly exceed the capacity of most health systems to supply them. Peruâ&#x20AC;&#x2122;s experience with its SIS exemplifies the challenges, which boil down to channelling a relatively large flow

Policy, Institutional and Methodological Lessons

307

of resources towards a large social need as effectively as possible. What is needed in this area is more analysis of how to make systems such as this one more effective, partly though better internal organization and partly through better understanding of which services and interventions bring the biggest social payoff. Some interventions have a notoriously high cost for a low payoff; others, presumably, are at the other end of the spectrum. vii) Emigration is an important safety valve for low-income families, as in Tajikistan now, Malawi earlier and Mexico albeit at a different level of earnings and development from the other two. Optimal social protection may be substantially different in such cases from what it is elsewhere—involving, for example, the special educational challenges of the children of emigrants and the special needs of low-income families without migrants. In the institutional area, points of interest include: i) the importance and effectiveness of community participation, as in Peru’s Provias road programme, including especially the participation of women, whose needs and priorities are otherwise less likely to be taken duly into account in project design and implementation; ii) the relationship between targeting goals and programme efficiency, as exemplified from Malawi by the advantages of not attempting precise targeting of start-up packages to the poorest families and the difficulty of forcing graduation from the Peruvian Glass of Milk programme and its continuity and possibly its efficiency at the delivery stage; and iii) the possibly enormous cost of lack of policy continuity where an approach has generated high positive momentum, as with the maize subsidy program in Malawi On the methodological front, useful lessons include the following: i) Especially in the case of large programmes such as India’s NREGA it is important to both undertake a general equilibrium analysis with the capacity to identify and quantify likely indirect effects and eventually test programme success against the evolution of aggregate variables. The latter can only be done where a policy has major scope, as with the NREGA in India and fertilizer policy in Malawi, where the case studies have undertaken such analysis; it might also be possible for microfinance in Peru. ii) Longer- as well as shorter-term effects need to be measured to make distinctions between good shorter- and longer-term social protection policy packages. The challenge here is to be able to analyse and evaluate well after a programme’s initiation. It is especially important in programmes which are by nature long-lasting, such as Peru’s SIS, and whose evolution is very hard to predict at the start.

308

Social protection, growth and employment

Notes 1 Economies of scale exist when an increase by a given percentage of all inputs into a production process leads to an output increase of a greater percentage. Diseconomies of scale exist when the output increase is smaller percentagewise than the input increase. 2 See, for example, Rosenzweig (1999). 3 Primary school enrolment is high, that of secondary school about 30 percent of the age cohort and that of tertiary education about 1 percent. 4 This view is broadly consistent with the conclusions of the most in-depth study thus far of informality in Latin America, that of Perry et al. (2007). 5 Such analysis is effectively precluded in countries where microfinance is not large enough to make it reasonable to expect its overall or aggregate effects to be identifiable in the context of the various other determinants of these variables. 6 This is often the case. As highlighted by Shaffer (1999), ways to save time and effort expended in domestic tasks may be the most important benefits from such projects. 7 As of 2002 the share of urban employment that was informal in the ILO sense (self-employed and workers in establishments of up to five workers, except professionals and technical workers plus domestic servants) was 41.5 percent of total employment, compared to 30â&#x20AC;&#x201C;33 percent in Chile and Costa Rica, over 50 percent in Venezuela and Ecuador and 63 percent in Bolivia. 8 Use of cash transfers is presently being discussed as an at least partial substitute for the current system of food and subsidized fuel, which costs nearly 1 percent of GDP but is poorly managed and extremely inefficient (New York Times, 2013: 8). A small-scale project to deposit government pensions and scholarship funds directly into bank accounts with a view to eliminating middlemen has also been launched. 9 This outcome parallels that reported for Malawi in Chapter 2. The maize boom during the middle of the last decade brought significant indirect benefits to poor families along with the direct ones.

Policy, Institutional and Methodological Lessons

309

References Economic Commission for Latin America and the Caribbean (ECLAC), Statistical Yearbook for Latin America and the Caribbean, ECLAC, Santiago de Chile, 2004. Gasparini, L., G. Cruces, L. Tornarolli and M. Marchionni, ‘A Turning Point? Recent Developments on Inequality in Latin America and the Caribbean’, Documento de Trabajo, No. 81, Universidad Nacional de la Plata, Centro de Estudios Distributivos, Laborales y Sociales, La Plata, 2009. Harris, J. and M. Todaro, ‘Migration, unemployment and development: a Two-sector analysis’, American Economic Review, 1970, Vol. 60, No. 1. International Labour Office (ILO), Employment, Incomes and Equality: A Strategy for Increasing Productive Employment in Kenya, ILO, Geneva, 1972. International Monetary Fund (IMF), Republic of Tajikistan: Poverty Reduction Strategy Paper, IMF, Washington, DC, 2010. Johnson, B.F. and P. Kilby, Agriculture and structural transformation: Economic strategies in late-developing countries, Oxford University Press, New York, 1975. Johnston, B.F. and J. Mellor, ‘The Role of Agriculture in Economic Development’, American Economic Review, 1961, Vol. 51, No. 4: 566–593. Knaul, F.M., E. González-Pier, O. Gómez-Dantés et al., ‘The quest for universal health coverage: achieving social protection for all in Mexico’, The Lancet, 16 Aug. 2012, http://dx.doi.org/10.1016/S0140-6736(12)61068-X (accessed 29 Jan. 2013). Levy, S., Good Intentions, Bad Outcomes: Social Policy, Informality and Economic Growth in Mexico, Brookings Institution Press, Washington, DC, 2008. Lipton, M., Land Reform in Developing Countries; property rights and property wrongs. Routledge, London, 2009. New York Times, “India Aims to Lift its Poor with Cash Transfers, but Some See Political Motives”, The New York, Times, 6 Jan. 2013: 8. Perry, G.E., W.F. Maloney, O. Arias, P. Fajnzylber, A. Mason and J. Saavedra-Chanduvi, Informality, Exit and Exclusion, World Bank, Washington, DC, 2007. Pritchett, L., ‘Where Has All the Education Gone?’, The World Bank Economic Review, 2001, Vol. 15, No. 3: 367–391.

310

Social protection, growth and employment

Rodríguez, J. and M. Higa, ‘Informalidad, empleo y productividad en Perú’, Desafíos Laborales En América Latina después dedos décadas de reformas estructurales: Bolivia, Paraguay y Perú, 1997–2008, J. Rodriguez and A. Berry (eds), Instituto de Estudios Peruanos, International Development Research Centre and Pontificia Universidad Católica del Perú, Lima, 2010. Rosenzweig, M.R., ‘Schooling, Economic Growth, and Aggregate Data’, Development, Duality, and the International Regime: Essays in Honor of Gustav Ranis, G. Saxonhouse and T.N. Srinivasan (eds), University of Michigan Press, Ann Arbor, MI, 1999. Shaffer, P., The Poverty Debate with Application to the Republic of Guinea, University of Sussex, Brighton, 1999. The Economist, ‘Money where Your Mouth Is’, 10 November 2012: 41–42. World Bank, ‘Peru Rural Road Program: Bringing Access to the Rural Poor’, IBRD Results, World Bank, Washington, DC, 2010.